Skip to content

YantrikDB

v0.8.0 Feature

This release adds 6 notable features for engineering teams evaluating rollout.

✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agent-memory ai-agents anthropic claude-code cognitive-memory database
+12 more
embeddings hnsw knowledge-graph llm llm-memory mcp memory persistent-memory python rust semantic-memory vector-db

Summary

AI summary

Updates no-default-features, additive, and mem0/Zep/Letta across a mixed release.

Changes in this release

Feature Medium

Adds "World's Best Memory System" program features

Adds "World's Best Memory System" program features

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Introduces additive schema additions via CREATE TABLE IF NOT EXISTS

Introduces additive schema additions via CREATE TABLE IF NOT EXISTS

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Exposes 12 new public API methods through PyO3

Exposes 12 new public API methods through PyO3

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Adds namespace_importance_stats, skill_outcomes, artifact_repair_audit schema entries

Adds namespace_importance_stats, skill_outcomes, artifact_repair_audit schema entries

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Implements continuous entity backfill and auto_relate co-occurrence edges in knowledge graph (+2.5% recall)

Implements continuous entity backfill and auto_relate co-occurrence edges in knowledge graph (+2.5% recall)

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Adds runnable CI-gated harnesses for scaling self-benchmark, latency SLO, and competitor evaluation

Adds runnable CI-gated harnesses for scaling self-benchmark, latency SLO, and competitor evaluation

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Introduces `session_digest()` one-call briefing and end-of-session auto-capture

Introduces `session_digest()` one-call briefing and end-of-session auto-capture

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Adds trust metadata (age / superseded) on recall hits

Adds trust metadata (age / superseded) on recall hits

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Implements `chain_head()` with exact chain-head semantics and structural-intent recall hints

Implements `chain_head()` with exact chain-head semantics and structural-intent recall hints

Source: llm_adapter@2026-06-11

Confidence: high

Feature Low

Adds `list_records` surface for record listing

Adds `list_records` surface for record listing

Source: llm_adapter@2026-06-11

Confidence: low

Feature Low

Surfaces `list_records` for listing records

Surfaces `list_records` for listing records

Source: granite4.1:30b@2026-06-11-audit

Confidence: low

Full changelog

Major feature release implementing the "World's Best Memory System" program — closing the write-rich / close-poor audit diagnosis. All changes additive and backward-compatible (schema additions via CREATE TABLE IF NOT EXISTS; new public API only). Published to PyPI and crates.io.

Ingest integrity

Write-time tool-call artifact sanitizer + corpus repair migration; saturation-gated importance calibration + usage-corrected importance; mega-blob splitting into linked atomic facts.

Autonomous hygiene (close every loop)

Conflicts stamped on recall + auto burn-down; bounded trigger backlog (TTL + over-cap eviction); durable skill outcomes; one run_maintenance_cycle() drives every pass on a host schedule.

Structural query

chain_head() exact chain-head semantics; structural-intent recall hints; list_records surfaced.

Boot sequence

session_digest() one-call briefing; end-of-session auto-capture; trust metadata (age / superseded) on recall hits.

Knowledge graph

Continuous entity backfill + auto_relate co-occurrence edges (graph-lift eval verdict: +2.5% recall on connected data).

Evidence

Runnable, CI-gated harnesses — scaling self-benchmark (precision@scale + latency SLO), graph-lift eval, competitor harness (mem0/Zep/Letta). README claims now cite them.

New schema (additive): namespace_importance_stats, skill_outcomes, artifact_repair_audit. 12 new methods exposed through PyO3. 1591 Rust tests + Python eval suite green; slim (--no-default-features) green.

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Track YantrikDB

Get notified when new releases ship.

Sign up free

About YantrikDB

All releases →

Related context

Earlier breaking changes

  • v0.9.3 `recall_demand` isolation is namespace‑keyed and disabled on encrypted databases, purging legacy rows via migration.
  • v0.7.20 `correct()` now mutates in place, preserving rid and adding revision history (BREAKING CHANGE).
  • v0.7.9 Pure-additive; existing engines keep English models on v0.1.0.

Beta — feedback welcome: [email protected]