Skip to content

Release history

hindsight releases

Hindsight: Agent Memory That Learns

All releases

45 shown

No immediate action
v0.8.5 Breaking risk

control-plane, Eve, recall, integrations, deps

Review required
v0.8.4 Breaking risk
Dependencies

retain, recall, api, deps, integrations

No immediate action
v0.8.3 Breaking risk

api, mcp, metrics, consolidation, retain

No immediate action
v0.8.2 New feature

Cursor CLI, Cline, API embeddings, Obsidian, Memory

No immediate action
v0.8.1 Mixed

Cursor, Haystack, OMO integrations + fixes

No immediate action
v0.8.0 Breaking risk

api, recall, consolidation, transfer, config

Review required
v0.7.2 Breaking risk
Auth Breaking upgrade

api control-plane retain consolidation integrations docs CI perf

No immediate action
v0.7.1 Mixed

control-plane, OpenClaw, retrieval, API, embeddings

Review required
v0.7.0 Mixed
Dependencies Breaking upgrade

Docs, Blog, Typescript‑client, API, Control‑plane

Upgrade now
v0.6.2 Security relevant
Dependencies

Dependency security bumps

v0.6.1 Security relevant patches CVE-2023-4863 patches CVE-2026-42208 patches CVE-2026-42271
Notable features
  • Add AlloyDB ScaNN vector index support
  • Optional read-only backend for recall queries
Full changelog

What's Changed

  • docs: add 0.6.0 changelog and release blog post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1458
  • chore: remove self-driving-agents CLI — moved to own repo by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1461
  • fix(retain): reduce memory pressure by clearing content references after use by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1455
  • fix(openclaw): mission semantics + retainQueue config whitelist by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1473
  • feat(claude-code): create-agent skill understands SDA layout by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1475
  • chore(deps): bump pg0-embedded to >=0.14.0 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1476
  • feat(openclaw): drop redundant before_agent_start + add debugPerfTiming by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1477
  • feat(control-plane): enrich bank dropdown with memory stats by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1479
  • feat(engine): optional read-only backend for recall queries by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1460
  • fix(claude-code): bootstrap 'mcp' Python package on first run by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1485
  • feat(perf): publish perf-test results to external dashboard by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1474
  • fix(packaging): hard-pin meta packages to matching hindsight-api-slim by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1486
  • docs(claude-code): document knowledge tools and subagent skill by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1487
  • ci(release): add --provenance to npm publish (required for n8n Verified) by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1491
  • chore: stabilize CI — docs-skill pre-commit hook + retain dict-mutation fix by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1490
  • fix(docker): chmod 755 /home/hindsight for --user UID:GID support by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1493
  • feat(cli): add --strategy flag to memory retain-files by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1494
  • blog: Your Claude Code Subagents Don't Share What They Learn by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1456
  • docs(cli): document --strategy flag for memory retain-files by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1499
  • docs(env): document HINDSIGHT_API_READ_DATABASE_URL in .env.example by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1496
  • ci: cosign-sign published GHCR images (manual workflow) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1495
  • ci: cosign-sign release images + document verification by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1502
  • Add AlloyDB ScaNN vector index support by @can1357 in https://github.com/vectorize-io/hindsight/pull/1459
  • fix(worker): probe pg_proc before calling optional schemas_with_pending_work() by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1503
  • docs(installation): bake custom models into image instead of PVC by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1504
  • follow-up to #1459: AlloyDB ScaNN docs + review-nit cleanups by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1506
  • feat: add map-type entity labels for structured entity extraction by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1505
  • fix(openclaw): pass enableKnowledgeTools through getPluginConfig by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1507
  • docs(entity-labels): document type="map" structured entity groups by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1508
  • perf(worker): scope progress-stats fanout via existing routine by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1509
  • fix(n8n): drop hindsight-client runtime dep, inline HTTP calls by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1513
  • fix(openclaw): write hooks.allowConversationAccess in setup wizard by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1514
  • fix(openclaw): reuse existing token + URL when re-running setup wizard by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1518
  • feat(claude-code): resolve git worktrees + explicit directory-bank mapping by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1520
  • fix(openclaw): backfill plugins.allow with hindsight-openclaw in setup wizard by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1521
  • fix(engine): remove multiplicative retry layers in fact extraction by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1516
  • docs: surface stable worker_id guidance and zombie-operation recovery by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1522
  • fix(daemon): replace os.fork() with subprocess.Popen to fix MPS on macOS by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1519
  • blog: n8n Workflows Are Stateless. Hindsight Makes Them Compound. by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1511
  • fix(reflect): read document metadata from retain params by @nepenth in https://github.com/vectorize-io/hindsight/pull/1523
  • fix(worker): propagate child error_message to failed batch_retain parent by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1527
  • fix(claude-code): use detail=metadata for agent_knowledge_list_pages by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1528
  • feat(cp): add optional access-key login for Control Plane by @ariel-ai-bot in https://github.com/vectorize-io/hindsight/pull/1530
  • feat: add z.ai as LLM provider by @Burgunthy in https://github.com/vectorize-io/hindsight/pull/1529
  • feat(openclaw): prepend session-context block to retained transcripts by @kryptt in https://github.com/vectorize-io/hindsight/pull/1439
  • fix(cp): improve access-key auth UX and harden middleware by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1533
  • docs(zai): document z.ai provider and add default model by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1532
  • chore: fix formatting in llm_wrapper.py to pass verify-generated-files by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1534
  • perf(locomo): scope CI to a 3-conversation curated subset by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1536
  • fix(typescript-client): derive CLIENT_VERSION via tsup define by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1540
  • feat(api): add litellmrouter provider for LLM fallback chains by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1537

New Contributors

  • @can1357 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1459
  • @nepenth made their first contribution in https://github.com/vectorize-io/hindsight/pull/1523
  • @ariel-ai-bot made their first contribution in https://github.com/vectorize-io/hindsight/pull/1530
  • @Burgunthy made their first contribution in https://github.com/vectorize-io/hindsight/pull/1529
  • @kryptt made their first contribution in https://github.com/vectorize-io/hindsight/pull/1439

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.6.0...v0.6.1

v0.6.0 Breaking risk
Notable features
  • Oracle 23ai database backend added as supported enterprise storage
  • Self‑driving agents (part 1) introduced
Full changelog

What's Changed

  • Add Pipecat voice AI persistent memory blog post by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1300
  • docs(guides): add Hindsight update guides by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1301
  • Fix broken link in CLI ARM64 guide by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1305
  • Oracle 23ai database backend by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1307
  • feat(agentcore): add hindsight-agentcore integration for Bedrock AgentCore Runtime by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/822
  • fix(agentcore): async-native client, task tracking, drop per-package CHANGELOGs by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1313
  • fix(oracle): restore PG query semantics and clean up migration chain by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1312
  • refactor(release): single integration table in generate_changelog by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1314
  • docs: add scoring & ranking deep dive to recall docs by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1317
  • feat: add SmolAgents integration with Hindsight memory tools by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/658
  • fix(oracle): restore exact v0.5.6 PG query shapes by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1321
  • fix(test): remove stale profile auto-create assertion from bank stats test by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1323
  • fix(embeddings): add allowed_openai_params for OpenAI-compatible embedding dimensions by @zwcf5200 in https://github.com/vectorize-io/hindsight/pull/1320
  • fix: redact database URLs in config logs by @xmh1011 in https://github.com/vectorize-io/hindsight/pull/1316
  • docs(mental-models): document tags?source=mental_models per #1296 by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1311
  • feat(python-client): expose retain_async in retain() and aretain() by @harryplusplus in https://github.com/vectorize-io/hindsight/pull/1306
  • fix(codex): fix encoding with PowerShell by @jerviscui in https://github.com/vectorize-io/hindsight/pull/1185
  • chore: remove stray files accidentally landed on main by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1326
  • fix(deps): cap cryptography <47 to fix SIGILL on ARM64 Apple Silicon by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1324
  • feat: self-driving agents (part1) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1302
  • chore(docs): sync version-0.5 docs from next by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1328
  • perf(db): eliminate ResultRow wrapping overhead for PostgreSQL by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1325
  • perf: add recall-with-observations & consolidation suites, split CI steps, fix locomo by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1333
  • docs(models): document litellm-sdk embeddings provider by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1336
  • feat(perf): add HTTP mode to recall benchmark by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1315
  • fix(consolidator): dedupe + ON CONFLICT for observation_sources INSERT by @youchi1 in https://github.com/vectorize-io/hindsight/pull/1340
  • fix(embed): include 'all' extras when spawning hindsight-api from sibling source by @youchi1 in https://github.com/vectorize-io/hindsight/pull/1341
  • fix(async-ops): atomically commit batch_retain parent and child rows by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1343
  • docs(python-sdk): document retain_async kwarg per #1306 by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1347
  • docs: add Oracle Database as supported enterprise storage by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1356
  • fix(llm): guard against null content from OpenAI-compatible providers by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1355
  • feat(oracle): unify migrations under Alembic with dialect dispatcher by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1330
  • fix(config): default openai-codex model to gpt-5.4 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1357
  • fix(codex): filter synthetic AGENTS startup messages by @voarsh2 in https://github.com/vectorize-io/hindsight/pull/1346
  • feat(self-driving-agents): add nemoclaw harness support by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1335
  • docs: add Pydantic Logfire guide for Hindsight tracing by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1339
  • Blog: Agent Memory in SmolAgents with Hindsight Tools by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1329
  • Add AWS AgentCore integration blog post by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1377
  • fix(embed): drop gpt-4o-mini fallback when hindsight-api import fails by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1363
  • feat(typescript-client): add AbortSignal support to all HindsightClient methods by @harryplusplus in https://github.com/vectorize-io/hindsight/pull/1198
  • feat(stats): add time_field toggle to memories-timeseries chart by @aliu-ronin in https://github.com/vectorize-io/hindsight/pull/1246
  • feat(opencode): share memory bank across git worktrees of the same repo by @isac322 in https://github.com/vectorize-io/hindsight/pull/1352
  • docs(self-driving-agents): document nemoclaw harness + --sandbox flag by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1367
  • Harden OpenAI-compatible JSON responses and ensure json_object prompt hint by @voarsh2 in https://github.com/vectorize-io/hindsight/pull/1368
  • fix(hindsight-api-slim): bump opentelemetry-{api,sdk,instrumentation,exporter} floors so PrometheusMetricReader 0.62b1 doesn't crash startup by @nikolay-bratanov in https://github.com/vectorize-io/hindsight/pull/1373
  • fix(oracle): update CHECK constraints in baseline to match PG schema by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1379
  • fix(api): include observations in per-document graph and counts by @youchi1 in https://github.com/vectorize-io/hindsight/pull/1374
  • fix(webhooks): route webhook endpoints through tenant-aware engine methods by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1388
  • fix: redirect daemon subprocess stdout/stderr on POSIX to prevent TUI corruption by @laoli-no1 in https://github.com/vectorize-io/hindsight/pull/1380
  • fix(retain): retain all documents in async multi-document batches by @voarsh2 in https://github.com/vectorize-io/hindsight/pull/1382
  • chore(embed): tidy detach-popen helper and close log fds in parent by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1418
  • codex: add configurable recall timeout for UserPromptSubmit hooks by @voarsh2 in https://github.com/vectorize-io/hindsight/pull/1399
  • fix(openclaw): add WeakSet registration guard keyed by API instance by @vernmic in https://github.com/vectorize-io/hindsight/pull/1409
  • chore(clients): regenerate clients for time_field timeseries param by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1420
  • fix(typescript-client): expose missing recall/reflect params by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1362
  • fix: clean up async batch retain test fixtures and comments by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1419
  • fix(api): recognize Pydantic aliases in unknown param middleware by @Tosko4 in https://github.com/vectorize-io/hindsight/pull/1417
  • fix(retain): defer memory_links → memory_units FKs to break cascade deadlock by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1398
  • fix(daemon): honor --host and HINDSIGHT_API_HOST in daemon mode by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1422
  • fix(recall): preserve original exception in recall_async error path by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1421
  • fix(worker): handle NotImplementedError from add_signal_handler on Windows by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1423
  • fix(self-driving-agents): fail fast when nemoclaw sandbox is missing by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1365
  • fix(mcp): expose tag_groups parameter on recall tool by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1424
  • chore(generated): regenerate openapi spec + clients post #1246 by @aliu-ronin in https://github.com/vectorize-io/hindsight/pull/1426
  • fix(hindsight-embed): use sysconfig to find scripts dir in daemon start by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1425
  • feat(anthropic): env-driven max_retries + default_headers knobs by @TuftyBruno in https://github.com/vectorize-io/hindsight/pull/1389
  • feat(self-driving-agents): add Claude Chat/Cowork harness by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1427
  • feat(self-driving-agents): add Hermes Agent harness support by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1431
  • fix(entity-resolver): stamp cooccurrences with event_date, not now() by @aliu-ronin in https://github.com/vectorize-io/hindsight/pull/1247
  • blog: Your Agent Harness Has Tools. It Still Needs Memory. by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1429
  • docs(guides): add framework memory integration guides by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1432
  • feat(n8n): add n8n community-node package for Hindsight memory by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1364
  • fix(n8n): add missing entry point, fix auth header, add execution tests by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1444
  • ci: stop re-running test workflow on pull_request_review by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1446
  • fix: repair 4 broken tests on main by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1437
  • fix(recall): inherit observation entities through source_memory_ids by @youchi1 in https://github.com/vectorize-io/hindsight/pull/1397
  • feat(dify): add Dify integration with Hindsight memory tools by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1434
  • feat(claude-code): add knowledge tools via Python MCP server by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1428
  • fix(dify): rename package to hindsight-dify by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1451
  • chore: LLM minimum acceptance tests with CI-managed model matrix by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1445
  • fix(search): correct vchord BM25 score direction by @liling in https://github.com/vectorize-io/hindsight/pull/1453
  • feat(claude-code): knowledge tools, subagents, and create-agent skill by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1450
  • docs(claude-code): update README for v0.6.0 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1457
  • fix: regenerate docs skill to sync with removed config by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1454

New Contributors

  • @jerviscui made their first contribution in https://github.com/vectorize-io/hindsight/pull/1185
  • @youchi1 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1340
  • @voarsh2 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1346
  • @nikolay-bratanov made their first contribution in https://github.com/vectorize-io/hindsight/pull/1373
  • @laoli-no1 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1380
  • @vernmic made their first contribution in https://github.com/vectorize-io/hindsight/pull/1409
  • @TuftyBruno made their first contribution in https://github.com/vectorize-io/hindsight/pull/1389
  • @liling made their first contribution in https://github.com/vectorize-io/hindsight/pull/1453

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.5.6...v0.6.0

v0.5.6 Maintenance

Minor fixes and improvements.

Full changelog

What's Changed

  • docs(release): changelog + blog for v0.5.5 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1297
  • ci: include linux arm64 CLI in release assets by @xmh1011 in https://github.com/vectorize-io/hindsight/pull/1298

New Contributors

  • @xmh1011 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1298

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.5.5...v0.5.6

v0.5.5 Breaking risk
Breaking changes
  • Removes delete_memory MCP tool (authorization bypass fix)
  • GET /banks/{bank_id}/profile no longer auto-creates banks
  • Removes strength field from causal relations
Security fixes
  • Removes delete_memory MCP tool to close authorization-bypass gap
Notable features
  • Per-operation slot reservations for worker tasks
  • Pipecat voice AI pipeline memory integration
  • Full Windows platform support
Full changelog

What's Changed

  • release: 0.5.4 changelog by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1195
  • fix: auto-confirm control plane install on first UI launch by @bwjoke in https://github.com/vectorize-io/hindsight/pull/1197
  • feat(worker): per-operation slot reservations by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1199
  • feat(perf): add system performance test runner by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1201
  • fix(perf): fix perf-test CI install and WorkerPoller kwarg by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1206
  • feat(db): configurable Postgres statement_timeout on pool connections by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1200
  • fix(worker): prevent child tasks from blocking parent execution by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1210
  • chore(perf): default CI perf-test scale to large by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1211
  • Blog: Hindsight 10,000 Stars Celebration by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1208
  • Fix: Update 10k stars blog post date to April 22 by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1213
  • feat(api): expose retry_count and next_retry_at on operation responses by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1188
  • feat(retain): expose processed_content_tokens on RetainResult by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1217
  • fix(claude-code): retain on SessionEnd even when retainEveryNTurns > 1 by @starbit-biostar in https://github.com/vectorize-io/hindsight/pull/1152
  • feat(claude-code): recall from additional banks alongside primary by @starbit-biostar in https://github.com/vectorize-io/hindsight/pull/1153
  • feat(perf): add LoComo benchmark as parallel CI job by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1223
  • fix(claude-code): prevent compaction from overwriting retained memories by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1222
  • fix(integrations): preserve raw UTF-8 in dynamically-derived bank_id by @Desko77 in https://github.com/vectorize-io/hindsight/pull/1141
  • docs(workers): document per-operation slot reservations (#1199) by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1207
  • fix(perf): locomo defaults — all conversations, wait-consolidation, pro model by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1224
  • fix(mcp): remove delete_memory tool to close authorization-bypass gap by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1228
  • fix(ops): expose processing/cancelled statuses through API and UI by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1231
  • feat(api): add exclude_parents filter to list operations by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1230
  • feat(api,ui): document chunks API, reprocess, and enhanced document detail by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1236
  • fix: correct docker command in 10k stars blog quickstart by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1237
  • blog: Your Agent Is Not Forgetful. It Was Never Given a Memory. by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1235
  • guides: add 20 general agent memory guides by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1233
  • fix(litellm): handle streaming responses in _store_conversation by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1239
  • fix(claude-code): handle list-content tool_results in transcript parsing by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1226
  • fix(stats): timeseries buckets should return tz-aware ISO by @aliu-ronin in https://github.com/vectorize-io/hindsight/pull/1245
  • fix(db): accept asyncpg-style URLs for external PostgreSQL by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1225
  • feat(embeddings): allow Cohere output dimensions via env var by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1249
  • fix(docs): escape curly braces in generated changelog entries by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1248
  • fix(tests): repair 9 regressions surfaced on main by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1251
  • Update integration author for ContextForge by @ooa-andera in https://github.com/vectorize-io/hindsight/pull/1254
  • feat(embed): full Windows support + prefer sibling hindsight-api over uvx by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1250
  • fix(llm): handle DeepSeek tool-call compatibility by @M1p0 in https://github.com/vectorize-io/hindsight/pull/1253
  • docs: add nodes_by_fact_type field to Document Response Format example by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1243
  • feat(embeddings): add HINDSIGHT_API_EMBEDDINGS_GEMINI_FORCE_IPV4 opt-in by @connorblack in https://github.com/vectorize-io/hindsight/pull/1241
  • docs(ops): document processing + cancelled statuses from #1231 by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1238
  • docs(blog): align retainEveryNTurns default with #1186 (10 → 3) by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1268
  • docs(sidecar): mirror HINDSIGHT_API_EMBEDDINGS_GEMINI_FORCE_IPV4 from #1241 by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1265
  • fix(embed): satisfy ty on Windows-only subprocess attrs by @grimmjoww in https://github.com/vectorize-io/hindsight/pull/1263
  • docs(configuration): mirror COHERE_OUTPUT_DIMENSIONS env var into sidecar reference by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1252
  • chore(deps): bump actions/github-script from 8 to 9 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/1025
  • fix(openclaw): stop silently skipping retention on default agent:main:main sessions by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1276
  • fix(consolidation): reduce memory fan-out during consolidation recall by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1277
  • Add blog post: Connect ChatGPT and Perplexity to Hindsight for Long-Term Memory by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1255
  • chore(lint): cover all npm packages with prettier in lint.sh by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1279
  • fix(mcp): report Hindsight's version in serverInfo, not FastMCP's by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1281
  • docs(integrations): add ChatGPT and Perplexity integration guides by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1280
  • docs(installation): document memory footprint and hardware requirements by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1282
  • feat(pipecat): add Pipecat voice AI pipeline memory integration by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/921
  • fix(llm): simplify JSON schemas for better Ollama and LLM compliance by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1292
  • fix(parsers): LlamaParse follow-up — reuse client, fix error mapping, add tests by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1293
  • fix(api): GET /banks/{bank_id}/profile no longer auto-creates the bank by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1287
  • fix(retain): drop strength from causal relations (fixes #1289) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1295
  • fix(llm): omit tool_choice="auto" and add deepseek as first-class provider by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1294
  • chore(deps): bump actions/upload-artifact from 4 to 7 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/1278
  • test(integration): add Hermes Agent embedded-mode smoke test by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1283
  • feat: mental-models List view + /tags?source=mental_models by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1296

New Contributors

  • @bwjoke made their first contribution in https://github.com/vectorize-io/hindsight/pull/1197
  • @starbit-biostar made their first contribution in https://github.com/vectorize-io/hindsight/pull/1152
  • @Desko77 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1141
  • @aliu-ronin made their first contribution in https://github.com/vectorize-io/hindsight/pull/1245
  • @M1p0 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1253
  • @connorblack made their first contribution in https://github.com/vectorize-io/hindsight/pull/1241

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.5.4...v0.5.5

v0.5.4 New feature
Notable features
  • Worker decommission and status commands
  • Document CRUD methods
  • Memory engine execution improvements
Full changelog

What's Changed

  • release: 0.5.3 notes and blog post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1126
  • blog: OpenAI Agents persistent memory with Hindsight by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1129
  • fix(openai-agents): review followup — docs, tests, polish by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1134
  • docs: fix HINDSIGHT_API_LLM_MAX_RETRIES default (10 → 3) by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1137
  • docs(sdk): add document CRUD methods to TypeScript client reference by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1132
  • fix(ai-sdk): align ReflectBasedOn types with OpenAPI spec by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1163
  • fix(api): populate items_count in list_operations by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1164
  • feat(admin): add decommission-workers and worker-status CLI commands by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1165
  • feat(claude-code): add {user_id} retainTags template variable by @soichisumi in https://github.com/vectorize-io/hindsight/pull/1161
  • fix(startup): downgrade LLM verify_connection failure to warning by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1166
  • feat: disable daemon idle timeout by default by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1162
  • blog: OpenCode persistent memory with Hindsight by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1172
  • guide batch, OpenClaw and Hermes memory by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1177
  • fix(worker): pass DeferOperation through MemoryEngine.execute_task by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1135
  • fix(engine): use ensure_ascii=False in json.dumps for LLM prompts by @harryplusplus in https://github.com/vectorize-io/hindsight/pull/1169
  • docs: add cover images for Apr 20 OpenClaw and Hermes guide batch by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1181
  • fix(worker): scan for active schemas before claiming by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1109
  • fix(mcp): route update_bank through config resolver, expose all fields by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1168
  • docs(admin-cli): document decommission-workers and worker-status by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1180
  • fix(reflect): honor reflect_mission identity framing in prompt builder by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1167
  • fix(retain): prevent duplicate memory units from delta upsert chunk issues by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1178
  • fix(alembic): merge divergent heads for v0.5.3 by @grimmjoww in https://github.com/vectorize-io/hindsight/pull/1149
  • docs(mcp): document update_bank config_updates and configurable fields by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1183
  • blog: add five agent memory articles by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1184
  • fix(opencode): lower retainEveryNTurns default from 10 to 3 by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1186
  • fix(reflect): prevent directive leakage on empty banks by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1190
  • fix: allow reflect-specific LLM config when default is disabled by @zwcf5200 in https://github.com/vectorize-io/hindsight/pull/1189
  • chore(deps): bump actions/upload-pages-artifact from 4 to 5 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/1170
  • feat(embeddings): make OpenAI-compatible batch size configurable (#1142) by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1143
  • fix(embedded): recover from daemon crashes transparently by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1193
  • fix(retain): preserve document created_at across upsert + UI edit flow by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1194
  • fix(reflect): scope delta mental model recall to new memories only by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1192

New Contributors

  • @soichisumi made their first contribution in https://github.com/vectorize-io/hindsight/pull/1161
  • @harryplusplus made their first contribution in https://github.com/vectorize-io/hindsight/pull/1169
  • @zwcf5200 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1189

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.5.3...v0.5.4

v0.5.3 Mixed
Notable features
  • OpenAI Agents SDK integration
  • Mental models UI revamp with staleness signal and history reflection
  • Named connection profiles (-p/--profile)
Full changelog

What's Changed

  • feat(openclaw): session-scoped document_id and structured per-message timestamp by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1075
  • release: 0.5.2 notes and blog post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1074
  • docs: add screenshots to 0.5.2 release post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1078
  • fix(control-plane): encode bank ids in URLs end-to-end by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1079
  • fix(reflect): forward mental model max_tokens to refresh by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1076
  • feat(cli): add named connection profiles (-p/--profile) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1080
  • feat(api): add tenant field and configurable allowlist to JSON logs by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1085
  • feat(paperclip): replace library with Paperclip plugin (v0.2.0) by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/934
  • docs: add /guides section with Hermes how-to guides by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1062
  • fix(worker): per-tenant fair rotation in claim_batch by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/1088
  • feat(mental-models): staleness signal + history reflect snapshot + UI revamp by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1089
  • fix: handle 'timestamp' field for file retain API by @christerence in https://github.com/vectorize-io/hindsight/pull/1092
  • fix: files/retain upload problems and orphaned retains by @christerence in https://github.com/vectorize-io/hindsight/pull/1091
  • fix(worker): make submit_task idempotent when payload already set by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1097
  • test(file-retain): regression for timestamp -> event_date mapping by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1096
  • fix(openclaw): resolve full symlink chain in isDirectExecution() by @D2758695161 in https://github.com/vectorize-io/hindsight/pull/1093
  • feat(control-plane): surface failed-consolidation count and drilldown by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1100
  • fix(openclaw): per-session retain stops overwriting prior turns by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1102
  • fix(ollama): add think=false to _call_ollama_native payload by @karl-8888 in https://github.com/vectorize-io/hindsight/pull/1099
  • fix(consolidation): prevent orphan observations when source deleted mid-consolidation by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1090
  • feat(recall): make budget mapping configurable per bank by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1106
  • feat(worker): DeferOperation exception for extension-driven requeue by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1105
  • feat(mental-models): structured-ops delta refresh + observation cleanup on upsert by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1101
  • fix(control-plane): clearer constellation recency legend & node tooltip by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1108
  • blog: Constellation View and Entity Co-occurrence Graph by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1103
  • docs: add /guides section with how-to guides and comparisons by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1110
  • feat: add OpenAI Agents SDK integration by @DK09876 in https://github.com/vectorize-io/hindsight/pull/842
  • fix(integrations): add encoding="utf-8" to transcript file reads by @tordf in https://github.com/vectorize-io/hindsight/pull/1119
  • fix(ai-sdk): align HindsightClient interface with [email protected] by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1118
  • fix(migrations): restore broken chain for v0.4.22 to v0.5.x upgrades by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1117
  • fix(jina-mlx): serialize Metal GPU ops to prevent SIGSEGV on concurrent reranker calls by @lkttle in https://github.com/vectorize-io/hindsight/pull/1113
  • fix: improve reranker error messages and add configurable TEI timeout by @octo-patch in https://github.com/vectorize-io/hindsight/pull/1115
  • fix(config): reduce default LLM max retries from 10 to 3 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1121
  • feat: add consolidation_max_memories_per_round config by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1123

New Contributors

  • @christerence made their first contribution in https://github.com/vectorize-io/hindsight/pull/1092
  • @D2758695161 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1093
  • @karl-8888 made their first contribution in https://github.com/vectorize-io/hindsight/pull/1099
  • @tordf made their first contribution in https://github.com/vectorize-io/hindsight/pull/1119
  • @lkttle made their first contribution in https://github.com/vectorize-io/hindsight/pull/1113

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.5.2...v0.5.3

v0.5.2 Mixed
Notable features
  • Added hindsight-architect skill
  • Control plane: revamped bank stats view and added co-occurrence graph visualization
  • Added recall controls to mental model trigger API
Full changelog

What's Changed

  • fix(openclaw): register agent hooks on every plugin entry invocation by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1029
  • chore(deps): bump softprops/action-gh-release from 2 to 3 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/1024
  • fix(embedded): add timeout to _cleanup lock acquisition by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1023
  • blog: How I Built Multi-User AI Memory into a Financial Product from Day One by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1030
  • docs: 0.5.1 changelog entry and release blog post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1032
  • fix(openclaw): stabilize session identity and skip operational turns by @Tosko4 in https://github.com/vectorize-io/hindsight/pull/987
  • feat(openclaw): retain as Anthropic-shaped JSON with tool_use/tool_result blocks by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1031
  • fix(opencode-plugin): correct session.messages response shape and update tests by @apnea in https://github.com/vectorize-io/hindsight/pull/993
  • fix(opencode): shared state across sessions and post-compaction retain by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1034
  • feat: add hindsight-architect skill by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1035
  • blog: Update OpenClaw post for v0.6.0/v0.6.2 by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1038
  • fix(retain): prevent IndexError on embeddings/facts length mismatch (#1037) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1047
  • feat(worker): log [PENDING_BREAKDOWN] bucketing pending rows by claim filter by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1050
  • fix(bank-template): align BankTemplateConfig with _CONFIGURABLE_FIELDS by @mrkhachaturov in https://github.com/vectorize-io/hindsight/pull/1044
  • feat(operations): expose task_payload and document_ids on async ops by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1049
  • docs: reframe observations as evidence-grounded consolidated knowledge by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1051
  • feat(api): add recall controls to mental model trigger by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1048
  • fix(clients): identifying User-Agent headers (#1041) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1052
  • docs(configuration): document HINDSIGHT_API_RETAIN_CHUNK_BATCH_SIZE by @mrkhachaturov in https://github.com/vectorize-io/hindsight/pull/1045
  • feat(openclaw): merge inline retain tags with defaults by @Aldoustheorchestrator in https://github.com/vectorize-io/hindsight/pull/948
  • test: stabilize flaky retain and load batch tests by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1053
  • blog: What's new in hindsight-openclaw 0.6 by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/1040
  • feat(control-plane): revamp bank stats view and modernize shared UI primitives by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1055
  • fix(openclaw): make identity skip filters config-aware for per-agent banking by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1054
  • fix(opencode): replace unconditional console.error with debugLog by @DK09876 in https://github.com/vectorize-io/hindsight/pull/1057
  • docs(opencode): drop npm install step, document Hindsight Cloud by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1056
  • fix(ts-sdk): re-export BankTemplate types from package root by @mrkhachaturov in https://github.com/vectorize-io/hindsight/pull/1063
  • fix(docs): regenerate bank-template-schema.json and guard drift by @mrkhachaturov in https://github.com/vectorize-io/hindsight/pull/1065
  • feat(docs): split template manifests into per-file manifest_file refe… by @mrkhachaturov in https://github.com/vectorize-io/hindsight/pull/1066
  • fix: wire consolidation retry budget to LLM call site by @r266-tech in https://github.com/vectorize-io/hindsight/pull/1064
  • chore(lint): share ruff/prettier config across integrations by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1072
  • fix(consolidation): tighten retry budget config handling and repair tests by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1073
  • feat(entities): add co-occurrence graph view in control plane by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/1058

New Contributors

  • @Tosko4 made their first contribution in https://github.com/vectorize-io/hindsight/pull/987
  • @apnea made their first contribution in https://github.com/vectorize-io/hindsight/pull/993

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.5.1...v0.5.2

v0.5.1 New feature
Notable features
  • Resilient startup and richer retain metadata
  • Session pattern filtering for ignore and stateless sessions
  • HINDSIGHT_API_DEFAULT_BANK_TEMPLATE environment variable
v0.5.0 Breaking risk
Breaking changes
  • Removes BFS and MPFP graph retrieval strategies
Security fixes
  • Bump litellm to >=1.83.0 (critical CVE fix)
  • Bump vite (high CVE fix)
Notable features
  • AutoGen integration for persistent agent memory
  • Constellation view with Pretext canvas rendering
  • OpenCode persistent memory plugin
v0.4.22 Mixed
⚠ Upgrade required
  • API will now warn on unknown request parameters via X-Ignored-Params header
Notable features
  • Code-review skill with project standards
  • ZeroEntropy reranker with configurable base URL
  • Document metadata exposure in API and control plane
Full changelog

What's Changed

  • docs: 0.4.21 release blog post and changelog by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/765
  • blog: What's New in Hindsight Cloud — Native OAuth for MCP Clients by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/731
  • Update MCP OAuth blog post date to March 30 by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/769
  • Fix blog homepage: Hindsight Cloud section always shows top 3 posts by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/770
  • Fix trailing commas in openclaw plugin manifest by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/774
  • fix(docs): use tools=[] in BaseMemory example by @DK09876 in https://github.com/vectorize-io/hindsight/pull/772
  • feat: add configurable base URL for ZeroEntropy reranker by @iskhakovt in https://github.com/vectorize-io/hindsight/pull/766
  • feat(codex): structured tool call retention from rollout files by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/778
  • docs: add ZeroEntropy base URL config to docs by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/779
  • fix(llamaindex): document_id, memory API, and ReAct trace fixes by @DK09876 in https://github.com/vectorize-io/hindsight/pull/777
  • SEO: add title and description to all integration pages by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/787
  • fix(codex): merge new settings on upgrade instead of skipping by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/780
  • fix(engine): classify first-person agent experiences as 'experience' fact type by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/775
  • fix(gemini): thought_signature read from wrong object and type in gemini 3.1+ tool calls by @ParadiseN1 in https://github.com/vectorize-io/hindsight/pull/785
  • fix(openclaw): defer heavy init to service.start() to avoid CLI slowdown by @kagura-agent in https://github.com/vectorize-io/hindsight/pull/764
  • fix(openclaw): warn log and tests for CLI mode waitForReady no-op by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/799
  • ci: support safe-to-test label for fork PRs by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/800
  • fix(tests): use random port for pg0 to avoid port conflicts by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/801
  • fix(http): recall endpoint drops metadata in response by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/803
  • Add /code-review skill for automated code quality checks by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/805
  • feat(api): warn on unknown request parameters via X-Ignored-Params header by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/802
  • ci: report safe-to-test CI results on PR by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/807
  • feat: add /code-review skill with project standards by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/806
  • feat: expose document_metadata in API and control plane by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/798
  • fix(claude-code): disable built-in tools to prevent MCP tool deferral in call_with_tools by @mkremnev in https://github.com/vectorize-io/hindsight/pull/784
  • feat: add optional LiteLLM SDK embedding output dimensions by @bullbo in https://github.com/vectorize-io/hindsight/pull/809
  • ci: trigger full CI on PR approval instead of safe-to-test label by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/813
  • What's New: multi-org support and credit transfers (March 29) by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/815
  • chore(deps): bump actions/deploy-pages from 4 to 5 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/763
  • chore(deps): bump azure/setup-helm from 4 to 5 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/761
  • chore(deps): bump dorny/paths-filter from 3 to 4 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/762
  • feat(api): add HINDSIGHT_API_LLM_EXTRA_BODY config for custom model params by @emirhan-gazi in https://github.com/vectorize-io/hindsight/pull/781
  • fix(reranker): use httpx for Cohere Azure endpoints to avoid 404 (fixes #783) by @kagura-agent in https://github.com/vectorize-io/hindsight/pull/790
  • refactor: replace LLMProvider classmethods with from_env() and document missing config fields by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/816
  • fix(mental-models): add tags_match and tag_groups to trigger config by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/804

New Contributors

  • @iskhakovt made their first contribution in https://github.com/vectorize-io/hindsight/pull/766
  • @ParadiseN1 made their first contribution in https://github.com/vectorize-io/hindsight/pull/785
  • @mkremnev made their first contribution in https://github.com/vectorize-io/hindsight/pull/784
  • @bullbo made their first contribution in https://github.com/vectorize-io/hindsight/pull/809
  • @emirhan-gazi made their first contribution in https://github.com/vectorize-io/hindsight/pull/781

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.21...v0.4.22

v0.4.21 Security relevant
Security fixes
  • LiteLLM 1.82.8 supply chain compromise
Notable features
  • Strands Agents SDK integration
  • LiteLLM provider for Bedrock and 100+ providers
  • Windows native support
v0.4.20 New feature
Notable features
  • LangGraph integration for persistent agent memory
  • Independent versioning for integrations
  • MCP tool improvements and filtering
v0.4.19 Mixed
Notable features
  • Agno and Hermes Agent integrations for memory persistence
  • Retain modes: verbatim, chunks, and named strategies
  • TypeScript client Deno compatibility
Full changelog

What's Changed

  • doc: add 0.4.18 release blog post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/567
  • fix: support for gemini-3.1-flash-lite-preview (thought_signature in tool calls) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/568
  • docs: improve type hints and documentation in client_wrapper by @hiSandog in https://github.com/vectorize-io/hindsight/pull/570
  • fix: inject Accept header in MCP middleware to prevent 406 errors by @DK09876 in https://github.com/vectorize-io/hindsight/pull/571
  • blog: add disposition-aware agents post by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/566
  • Upgrade GitHub Actions to latest versions by @salmanmkc in https://github.com/vectorize-io/hindsight/pull/576
  • chore(deps): bump actions/setup-python from 5 to 6 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/583
  • chore(deps): bump actions/download-artifact from 4 to 8 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/582
  • chore(deps): bump actions/checkout from 4 to 6 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/581
  • fix: chunk FK cascade so doc deletion doesn't leave orphan memory units by @jnMetaCode in https://github.com/vectorize-io/hindsight/pull/580
  • fix(migration): backsweep orphaned observation memory units by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/584
  • feat: local reranker FP16, bucket batching, and transformers 5.x compatibility by @fabioscarsi in https://github.com/vectorize-io/hindsight/pull/588
  • fix(docker): honor HINDSIGHT_CP_HOSTNAME for control-plane startup by @abix5 in https://github.com/vectorize-io/hindsight/pull/590
  • docs: add config vars for local reranker FP16 and bucket batching by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/589
  • docs(skills): encourage rich context over pre-summarized strings in retain by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/594
  • blog: add n8n persistent memory workflows post by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/585
  • docs: add Best Practices unversioned page by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/598
  • feat: hindsight-hermes integration for Hermes Agent by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/600
  • feat(retain): verbatim, chunks modes and named retain strategies by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/593
  • blog: Streamlit chatbot with persistent memory by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/602
  • fix: prevent silent memory loss on consolidation LLM failure by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/601
  • blog: fix streamlit post slug and add cover image by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/604
  • blog: fix internal links in streamlit post by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/605
  • feat: add Agno integration with Hindsight memory toolkit by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/596
  • feat(typescript-client): Deno compatibility by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/607

New Contributors

  • @hiSandog made their first contribution in https://github.com/vectorize-io/hindsight/pull/570
  • @jnMetaCode made their first contribution in https://github.com/vectorize-io/hindsight/pull/580

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.18...v0.4.19

v0.4.18 Mixed
Notable features
  • MiniMax LLM and Jina-mlx reranker for Apple Silicon support
  • Compound tag filtering via tag_groups
  • Configurable recall max query tokens via environment variable
Full changelog

What's Changed

  • doc: add 0.4.17 release blog post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/538
  • perf: replace window-function retrieval with UNION ALL + per-bank HNSW indexes by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/541
  • feat: make recall max query tokens configurable via env var by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/544
  • feat: add jina-mlx reranker provider for Apple Silicon by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/542
  • doc: Run Hindsight with Ollama: Local AI Memory, No API Keys Needed by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/536
  • doc: What's New in Hindsight Cloud — Programmatic API Key Management by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/543
  • fix: cancel in-flight async ops when bank is deleted by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/545
  • fix: register embedded profiles in CLI metadata on daemon start by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/546
  • fix(openclaw): keep recalled memories out of the visible transcript by @stablegenius49 in https://github.com/vectorize-io/hindsight/pull/548
  • blog: Time-Aware Spreading Activation for Memory Graphs by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/547
  • feat: add MiniMax LLM provider support by @octo-patch in https://github.com/vectorize-io/hindsight/pull/550
  • fix: truncate documents exceeding LiteLLM reranker context limit by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/549
  • Upgrade GitHub Actions to latest versions by @salmanmkc in https://github.com/vectorize-io/hindsight/pull/553
  • chore: add dependabot config for GitHub Actions updates by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/554
  • chore(deps): bump actions/cache from 4 to 5 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/559
  • chore(deps): bump actions/setup-go from 5 to 6 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/558
  • chore(deps): bump actions/upload-artifact from 4 to 7 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/555
  • chore(deps): bump actions/checkout from 4 to 6 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/556
  • feat: introduce hindsight-api-slim and hindsight-all-slim packages by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/560
  • fix: remove broken minimax test and enhance slim smoke tests by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/564
  • chore(deps): bump actions/setup-node from 4 to 6 by @dependabot[bot] in https://github.com/vectorize-io/hindsight/pull/557
  • docs: revamp sidebar with icon grids and language support by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/563
  • feat: compound tag filtering via tag_groups by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/562

New Contributors

  • @stablegenius49 made their first contribution in https://github.com/vectorize-io/hindsight/pull/548
  • @octo-patch made their first contribution in https://github.com/vectorize-io/hindsight/pull/550
  • @salmanmkc made their first contribution in https://github.com/vectorize-io/hindsight/pull/553
  • @dependabot[bot] made their first contribution in https://github.com/vectorize-io/hindsight/pull/559

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.17...v0.4.18

v0.4.17 New feature
Notable features
  • Per-request file parser selection with fallback chains
  • Observation history tracking with diff UI
  • Mental model history tracking and UI diff view
v0.4.16 New feature
Notable features
  • Webhook system with retain.completed event
  • OpenClaw v2 recall and retention controls
  • Performance improvements with GIN indexing
v0.4.15 Mixed
Notable features
  • Pydantic AI integration for persistent agent memory
  • observation_scopes field to control observation granularity
  • Entity labels with optional values and multi-value support
Full changelog

What's Changed

  • doc: changelog and blog post by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/445
  • doc: improvements by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/448
  • doc: 0.4.14 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/449
  • feat: observation_scopes field to drive observations granularity by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/447
  • feat(openclaw): retain last n+2 turns every n turns (default n=10) by @fabioscarsi in https://github.com/vectorize-io/hindsight/pull/452
  • Add bank-scoped validation to engine and HTTP handlers by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/454
  • fix: JSON serialization and logging exception propagation in claude_code_llm by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/461
  • fix: zeroentropy rerank URL missing /v1 prefix and MCP retain async_processing param by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/460
  • fix(control-plane): observations count always showing 0 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/464
  • fix: resolve consolidation deadlock caused by zombie processing tasks on retry by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/463
  • fix(reflect): prevent context_length_exceeded on large memory banks by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/462
  • feat: support timestamp="unset" to retain content without a date by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/465
  • feat: entity labels — optional, free_values, multi_value, UI polish by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/450
  • docs: entities vs tags vs metadata by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/466
  • feat: add Pydantic AI integration for persistent agent memory by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/441
  • feat: add Pydantic AI integration to CI, release pipeline, and docs by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/467
  • feat: add tags filtering and q description fix for list documents API by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/468
  • blog: add CrewAI persistent memory post by @benfrank241 in https://github.com/vectorize-io/hindsight/pull/471
  • feat: add extension hooks for root routing and error headers by @DK09876 in https://github.com/vectorize-io/hindsight/pull/470
  • refactor(openclaw): replace console.log with debug() helper by @slayoffer in https://github.com/vectorize-io/hindsight/pull/456
  • fix(performance): improve recall and retain performance on large banks by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/469
  • feat: configurable Gemini/Vertex AI safety settings by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/473
  • perf(recall): improve single-query chunk fetch by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/475
  • fix(ts-sdk): send null instead of undefined when includeEntities is false by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/476
  • refactor: replace set_gemini_safety_settings() with LLMProvider.with_config() by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/474

New Contributors

  • @fabioscarsi made their first contribution in https://github.com/vectorize-io/hindsight/pull/452

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.14...v0.4.15

v0.4.14 New feature
Notable features
  • Chat SDK integration for persistent chatbot memory
  • ZeroEntropy reranker support
  • Batch observations consolidation
v0.4.13 New feature
Notable features
  • CrewAI integration for persistent crew memory
  • Improved temporal ordering precision
  • Fixed docker startup with named volumes
v0.4.12 Mixed
Security fixes
  • Shell safety hardening in openclaw
Notable features
  • Go client SDK with ogen code generation
  • Support for PDF, images, and Office documents
  • Batch API support for retain with OpenAI and Groq
Full changelog

What's Changed

  • doc: changelog and blog post for 0.4.11 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/363
  • feat: allow chunks-only in recall (max_tokens=0) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/364
  • fix(openclaw): remove unused imports, retry health check, suppress unhandled rejection by @slayoffer in https://github.com/vectorize-io/hindsight/pull/373
  • fix: propagate document_tags in async retain path by @abix5 in https://github.com/vectorize-io/hindsight/pull/374
  • feat: add Go client SDK with ogen code generation by @franchb in https://github.com/vectorize-io/hindsight/pull/375
  • fix: improve async batch retain with large payloads by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/366
  • feat: support Batch API for retain (openai/groq) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/365
  • feat: use official go generator for Go client by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/377
  • feat: support for pgvectorscale (DiskANN) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/378
  • doc: add go client examples by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/380
  • ci: ensure docs get created with no extracted facts by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/379
  • feat: support azure pg_diskann by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/381
  • doc: add faq page by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/383
  • fix(openclaw): error E2BIG on large content ingested by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/389
  • fix(python-client): async method parity and server keepalive timeout by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/387
  • feat: accept pdf, images and office files by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/390
  • fix: restore entity retrieval in recall by @dcbouius in https://github.com/vectorize-io/hindsight/pull/391
  • fix(go-client): use monorepo-compatible module path by @franchb in https://github.com/vectorize-io/hindsight/pull/392
  • feat: improve ai sdk tools by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/394
  • fix(go-client): add go build to CI by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/393
  • fix(openclaw): shell safety, HTTP dual-mode, lazy reinit, per-user banks by @slayoffer in https://github.com/vectorize-io/hindsight/pull/388
  • feat: add iris as file parser by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/395
  • fix: improve openclaw test coverage by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/396

New Contributors

  • @abix5 made their first contribution in https://github.com/vectorize-io/hindsight/pull/374
  • @franchb made their first contribution in https://github.com/vectorize-io/hindsight/pull/375

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.11...v0.4.12

v0.4.11 Mixed
Notable features
  • Hierarchical configuration system supporting system, tenant, and bank-level settings
  • Vector database support (pgvector, vchord) with PostgreSQL text/vector extensions including timescale pg_textsearch and litellm-sdk reranker integration
  • MCP server enhancements: mental model CRUD tools, hardened routing/validation, reverse proxy support, improved usage metering with tenant isolation
Full changelog

What's Changed

  • fix(helm): gke overriding HINDSIGHT_API_PORT by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/328
  • feat(helm): add PDB and per-component affinity support by @nuclon in https://github.com/vectorize-io/hindsight/pull/327
  • feat(openclaw): add excludeProviders config by @GodsBoy in https://github.com/vectorize-io/hindsight/pull/332
  • Fix MCP usage metering: propagate tenant_id through context vars by @DK09876 in https://github.com/vectorize-io/hindsight/pull/334
  • feat: add otel traceability by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/330
  • feat: add docs skill by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/335
  • fix: include tiktoken in slim image by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/336
  • Fix async batch retain incorrectly marked as internal by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/338
  • feat: add mental model CRUD tools to MCP server by @DK09876 in https://github.com/vectorize-io/hindsight/pull/337
  • feat(helm): add TEI reranker sidecar support by @slayoffer in https://github.com/vectorize-io/hindsight/pull/333
  • Harden MCP server: fix routing, validation, and usage metering by @DK09876 in https://github.com/vectorize-io/hindsight/pull/341
  • Add actual LLM token usage fields to RetainResult by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/342
  • fix: improve model configuration for litellm gateway by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/345
  • fix: add trust_code env config by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/347
  • Replace waitlist links with direct signup URL by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/349
  • Fix MCP extra args rejection and bank ID resolution priority by @DK09876 in https://github.com/vectorize-io/hindsight/pull/351
  • feat: add reverse proxy support by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/346
  • Implement the vchord / pgvector support by @qdrddr in https://github.com/vectorize-io/hindsight/pull/350
  • fix: resolve based_on schema/serialization issues in reflect API by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/348
  • feat: implement hierarchical configuration (system, tenant, bank) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/329
  • feat: support for other text and vector search pg extensions by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/355
  • chore: remove dead code by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/356
  • Fix incorrect MCP tool parameters in docs by @DK09876 in https://github.com/vectorize-io/hindsight/pull/358
  • feat: support timescale pg_textsearch as text search extension by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/359
  • feat: support litellm-sdk as reranker and embeddings by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/357
  • fix(openclaw): avoid memory retain recursion by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/362

New Contributors

  • @qdrddr made their first contribution in https://github.com/vectorize-io/hindsight/pull/350

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.10...v0.4.11

v0.4.10 Mixed
Security fixes
  • Prevents database credentials from being logged
Notable features
  • Markdown support in reflect and mental models
  • TenantExtension authentication for MCP endpoints
  • Supabase tenant extension
Full changelog

What's Changed

  • fixed cast error by @haydenrear in https://github.com/vectorize-io/hindsight/pull/300
  • fix: tagged directives should be applied to tagged mental models by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/303
  • docs: add AI SDK integration documentation by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/304
  • ci: ensure backwards/forward compatibility of the API by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/306
  • fix(openclaw): remove format:uri to fix ajv warning by @GodsBoy in https://github.com/vectorize-io/hindsight/pull/309
  • feat: support markdown in reflect and mental models by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/307
  • ci: ensure python 3.14 compatibility by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/310
  • fix(ci): resolve flaky test failures in api tests by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/311
  • feat: slim docker distro by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/314
  • doc: update claude-code usage terms by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/315
  • fix: hindsight-embed profiles are not loaded correctly by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/316
  • feat: add TenantExtension auth to MCP endpoint by @DK09876 in https://github.com/vectorize-io/hindsight/pull/286
  • doc: improve Node.js client example by @vanvuongngo in https://github.com/vectorize-io/hindsight/pull/320
  • feat: improve mcp tools based on endpoint by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/318
  • fix(openclaw): prevent memory wipe on every session by @slayoffer in https://github.com/vectorize-io/hindsight/pull/323
  • feat: add docker-compose example by @vanvuongngo in https://github.com/vectorize-io/hindsight/pull/313
  • fix: do not log db user/password by @vanvuongngo in https://github.com/vectorize-io/hindsight/pull/312
  • Add Supabase tenant extension by @jerryhenley in https://github.com/vectorize-io/hindsight/pull/267
  • fix(helm): improve appVersion usage by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/326
  • doc: prepare doc for 0.4.10 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/325

New Contributors

  • @haydenrear made their first contribution in https://github.com/vectorize-io/hindsight/pull/300
  • @vanvuongngo made their first contribution in https://github.com/vectorize-io/hindsight/pull/320
  • @jerryhenley made their first contribution in https://github.com/vectorize-io/hindsight/pull/267

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.9...v0.4.10

v0.4.9 Mixed
Notable features
  • HindsightEmbedded Python SDK for programmatic Hindsight access
  • External Hindsight API support with dynamic per-channel memory banks for OpenClaw
  • Streaming support on hindsight-litellm wrappers and AI SDK integration
Full changelog

What's Changed

  • doc: changelog for 0.4.8 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/283
  • doc: update cookbook by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/284
  • fix(openclaw): improve shell argument escaping by @slayoffer in https://github.com/vectorize-io/hindsight/pull/288
  • feat(openclaw): add external Hindsight API support by @slayoffer in https://github.com/vectorize-io/hindsight/pull/289
  • docs: expand external API configuration for OpenClaw by @slayoffer in https://github.com/vectorize-io/hindsight/pull/294
  • feat(openclaw): add dynamic per-channel memory banks by @slayoffer in https://github.com/vectorize-io/hindsight/pull/290
  • fix: hide hf logging by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/295
  • fix: improve claude code and codex for /reflect by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/285
  • feat(hindsight-litellm): support streaming on wrappers by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/296
  • feat: HindsightEmbedded python SDK by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/293
  • feat: improve mental models ux on control plane by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/297
  • Fix recall endpoint timeout handling and add query length validation by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/298
  • feat: ai sdk integration by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/299

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.8...v0.4.9

v0.4.8 Security relevant
⚠ Upgrade required
  • Fixed unreliable custom PostgreSQL schema support
  • Request context now propagated through async task payloads
Security fixes
  • Upgraded vulnerable dependencies
Notable features
  • Support for Codex and Claude Code as LLM providers
  • New hindisght-embed profiles with enhanced UX via rich logging and profile isolation
  • Pre-operation validation hooks for mental model create and refresh
Full changelog

What's Changed

  • Add pre-operation validation hooks for mental model create/refresh by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/271
  • Propagate request context through async task payloads by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/273
  • feat(openclaw): add llmProvider/llmModel plugin config options by @GodsBoy in https://github.com/vectorize-io/hindsight/pull/274
  • feat: print version during startup by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/275
  • feat: support for codex and claude-code as llm by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/276
  • feat(embed): add hindisght-embed profiles by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/277
  • fix: custom pg schema is not reliable by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/278
  • Fix: load operation validator extension in worker by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/280
  • feat: improve openclaw and hindisght-embed params by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/279
  • fix(sec): upgrade vulnerable deps by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/254
  • fix: improve embed ux with rich logging and profile isolation by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/282

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.7...v0.4.8

v0.4.7 Mixed
Security fixes
  • Sanitize null bytes from text fields before PostgreSQL insertion
Notable features
  • Docker images now preload tiktoken encoding during build
  • Extension hooks for mental model operations
  • Hindsight-embed external API support and OpenClaw fixes
Full changelog

What's Changed

  • fix: worker doesn't pick up correct default schema by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/259
  • fix: sanitize null bytes from text fields before PostgreSQL insertion by @slayoffer in https://github.com/vectorize-io/hindsight/pull/238
  • feat(docker): preload tiktoken encoding during build by @slayoffer in https://github.com/vectorize-io/hindsight/pull/249
  • fix(hindsight-embed): respect HINDSIGHT_API_DATABASE_URL if already set by @GodsBoy in https://github.com/vectorize-io/hindsight/pull/262
  • Add extension hooks for mental model operations by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/260
  • feat(hindsight-embed): external API support + OpenClaw fixes (#263, #264) by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/265

New Contributors

  • @GodsBoy made their first contribution in https://github.com/vectorize-io/hindsight/pull/262

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.6...v0.4.7

v0.4.4 Bug fix

## What's Changed * fix: retain async fails if timestamp is set * fix: rename openclawd to openclaw

v0.4.3 Breaking risk
⚠ Upgrade required
  • Update code references from moltbot to openclawd
  • Tenant schema migrations now run automatically at startup
Breaking changes
  • moltbot renamed to openclawd
Notable features
  • Vertex AI now supported as LLM provider
  • CLI: --wait flag for consolidate command and --date filter for document list
  • MCP: Bearer token authentication and tenant auth propagation
Full changelog

What's Changed

  • fix: pass tenant extension to worker MemoryEngine for correct schema context by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/236
  • fix: run migrations on tenant schemas at startup and harden worker poller by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/237
  • feat: support vertex as llm provider by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/233
  • Switch Vertex AI provider to native genai SDK by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/242
  • fix(control-plane): handle undefined response.data in graph route by @slayoffer in https://github.com/vectorize-io/hindsight/pull/239
  • feat(cli): add --wait flag for consolidate and --date filter for document list by @slayoffer in https://github.com/vectorize-io/hindsight/pull/244
  • fix(control-plane): pass API key to dataplane for tenant auth by @slayoffer in https://github.com/vectorize-io/hindsight/pull/243
  • fix(auth): skip tenant auth for all internal background tasks by @slayoffer in https://github.com/vectorize-io/hindsight/pull/240
  • feat(mcp): add Bearer token authentication and tenant auth propagation by @slayoffer in https://github.com/vectorize-io/hindsight/pull/241
  • chore: remove dead code by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/245
  • fix: improve doc on vertexai and mcp by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/247
  • fix(docker): add retry logic for ML model downloads by @slayoffer in https://github.com/vectorize-io/hindsight/pull/248
  • fix: rename moltbot to openclawd by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/246
  • fix: deadlock in worker polling by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/250

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.2...v0.4.3

v0.4.2 Mixed
Notable features
  • moltbot integration
  • Real-time timing breakdown logging for consolidation
  • Expanded LLM retry configuration options
Full changelog

What's Changed

  • fix(doc): improve docs versioning and release by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/231
  • fix: hindsight-embed on macos crashes by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/228
  • feat: moltbot integration by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/216
  • feat: add real-time timing breakdown logging for consolidation by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/235
  • feat: add more config options for llm retries by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/234

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.4.1...v0.4.2

v0.4.0 Security relevant
Security fixes
  • Vulnerable dependencies upgraded
Notable features
  • Mental models framework with directives and reflections API
  • Cloud mode in skill installer for team memory sharing
  • New worker service for distributed task execution
Full changelog

What's Changed

  • doc: changelog for 0.3.0 by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/156
  • doc: update expired Slack invite link by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/157
  • feat: add cloud mode to skill installer for team memory sharing by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/158
  • Fix skill installer test examples to use meaningful content by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/160
  • feat(cli): accept more file types on retain-files by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/163
  • feat: introduce mental models by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/132
  • doc: refinement for 0.3.0 new features by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/159
  • Add structured JSON logging support by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/170
  • feat: improve mental model refresh and add directives by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/166
  • chore: unify agents.md and claude.md by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/173
  • fix(sec): upgrade vulnerable deps by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/174
  • feat(clients): mental models api by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/172
  • fix: pytorch init failures by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/175
  • feat: new 'worker' service by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/176
  • chore: drop unused access_count column by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/178
  • fix: improve pytorch model initialization to prevent meta tensor issues by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/180
  • fix: sometimes memories gets extracted in the wrong language by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/184
  • fix: simplify pytorch model initialization to prevent meta tensor issues by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/185
  • feat: revisit mental models, directives and reflections by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/179
  • Fix Gemini tool response format by including function name by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/187
  • feat(python-sdk): add tags filtering support to high-level client by @phamgialinhlx in https://github.com/vectorize-io/hindsight/pull/186
  • feat: support for npx add-skill by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/191
  • feat(mcp): add timestamp to retain by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/190
  • fix: improve mental model consolidation by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/197
  • feat(litellm): async retain, reflect support, and API cleanup by @DK09876 in https://github.com/vectorize-io/hindsight/pull/167
  • Fix: Pass api_key to Hindsight client in litellm integration by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/193
  • chore: versioned docs by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/198
  • doc: mental models by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/199
  • ci: add upgrade tests by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/200
  • fix(ui): reflections based on don't show up all contents by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/203
  • feat(litellm): support tags and mission in litellm package by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/202
  • chore: internal renames by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/204
  • fix: include tags, created_at, proof_count in graph table_rows by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/207
  • fix: multi-tenant schema context for worker task execution by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/208
  • chore: drop dead code by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/210
  • fix: misc fixes for observations and mental models by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/209
  • chore: cleanup benchmarks runner with old flags by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/212
  • feat: add custom extraction prompt by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/213
  • fix: graph endpoint not showing links for observations by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/214
  • fix(embed): daemon process XPC connection crash on macos by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/215

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.3.0...v0.4.0

v0.3.0 Security relevant
Security fixes
  • CVE-2025-15284 in qs
Notable features
  • Configurable embedding dimensions with OpenAI Embeddings support
  • Backup and restore functionality
  • Offline database migration support
v0.2.1 Maintenance

## What's Changed * doc: changelog for 0.2.0 (and regenerate clients)

v0.2.0 Mixed
Notable features
  • Anthropic Claude and LM Studio provider support with local LLM improvements for reasoning models
  • Structured output and configurable max_tokens in /reflect endpoint
  • Multi-bank MCP access with new MCP tools and graph visualization
Full changelog

What's Changed

  • doc: add skills documentation by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/73
  • Load operation validator extension in main entry point by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/72
  • feat(config): Add configurable observation thresholds by @bjornslib in https://github.com/vectorize-io/hindsight/pull/83
  • fix(mcp): Chain MCP lifespan with FastAPI app lifespan by @bjornslib in https://github.com/vectorize-io/hindsight/pull/81
  • feat: Add Anthropic Claude and LM Studio provider support by @csfet9 in https://github.com/vectorize-io/hindsight/pull/36
  • feat(doc): add new config options and supported providers by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/84
  • feat: add max_tokens and structured output to /reflect by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/74
  • Add operation validator extension support with proper HTTP error handling by @cdbartholomew in https://github.com/vectorize-io/hindsight/pull/86
  • Feature/graph viz by @chrislatimer in https://github.com/vectorize-io/hindsight/pull/85
  • feat: Add local LLM improvements for reasoning models and Docker startup by @csfet9 in https://github.com/vectorize-io/hindsight/pull/88
  • feat: Add user-provided entities support to retain endpoint by @phamgialinhlx in https://github.com/vectorize-io/hindsight/pull/91
  • feat(mcp): Add multi-bank access and new MCP tools by @bjornslib in https://github.com/vectorize-io/hindsight/pull/82
  • misc: add mcp integration tests and increase test coverage by @nicoloboschi in https://github.com/vectorize-io/hindsight/pull/98

New Contributors

  • @bjornslib made their first contribution in https://github.com/vectorize-io/hindsight/pull/83
  • @csfet9 made their first contribution in https://github.com/vectorize-io/hindsight/pull/36
  • @phamgialinhlx made their first contribution in https://github.com/vectorize-io/hindsight/pull/91

Full Changelog: https://github.com/vectorize-io/hindsight/compare/v0.1.16...v0.2.0

Beta — feedback welcome: [email protected]