✓ No known CVEs patched in this version
Topics
+12 more
Summary
AI summaryUpdates span web UI, seeder CLI, agent feedback, monitors limits, batch-export handling, and ClickHouse compatibility detection.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Low |
Adds experimental feature modal for web UI. Adds experimental feature modal for web UI. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Adds agent‑first seed CLI with doctor, complex trees/sessions/bulk, v4 mirroring. Adds agent‑first seed CLI with doctor, complex trees/sessions/bulk, v4 mirroring. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Detects ClickHouse versions on start and sets compatibility flags automatically. Detects ClickHouse versions on start and sets compatibility flags automatically. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Adds conversation starters to agent feature. Adds conversation starters to agent feature. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Buckets live preview by evaluation window in monitors. Buckets live preview by evaluation window in monitors. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Retires feature flag for monitors and raises organization limit to 20. Retires feature flag for monitors and raises organization limit to 20. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Updates support severity dropdown options. Updates support severity dropdown options. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Changes icon for LLM‑as‑a‑Judge evaluator setup in evals. Changes icon for LLM‑as‑a‑Judge evaluator setup in evals. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Removes beta label from corrections feature. Removes beta label from corrections feature. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Reads sessions from events table for v4 users in batch-export. Reads sessions from events table for v4 users in batch-export. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Feature | Low |
Gates enriched blob export availability on V4 environment variables for self‑hosted deployments. Gates enriched blob export availability on V4 environment variables for self‑hosted deployments. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Feature | Low |
Improves performance of checkboxes in traces/observations and adds shift‑selection support. Improves performance of checkboxes in traces/observations and adds shift‑selection support. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Bugfix | Medium |
Drops media link foreign keys in media module. Drops media link foreign keys in media module. Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Bugfix | Low |
Throttles Mixpanel integration exports in worker component. Throttles Mixpanel integration exports in worker component. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Bugfix | Low |
Resolves app theme for CodeBlock syntax highlighting in dark mode UI. Resolves app theme for CodeBlock syntax highlighting in dark mode UI. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Bugfix | Low |
Gates sessions events session_id pushdown on 'any of' operator logic. Gates sessions events session_id pushdown on 'any of' operator logic. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Bugfix | Low |
Snapshots v4 events‑table flag for session batch actions workflow. Snapshots v4 events‑table flag for session batch actions workflow. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Bugfix | Low |
Aligns agentic seed summaries in seeder tool. Aligns agentic seed summaries in seeder tool. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Bugfix | Low |
Reads scores trace metadata from events table for v4 users in batch-export. Reads scores trace metadata from events table for v4 users in batch-export. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
| Bugfix | Low |
Refines API error logging and prompt validation handling. Refines API error logging and prompt validation handling. Source: granite4.1:30b@2026-06-12-audit Confidence: low |
— |
Full changelog
Features
- feat(web): New experimental feature modal by @bezbac in https://github.com/langfuse/langfuse/pull/14152
- feat(seeder): agent-first seed CLI — doctor, complex trees/sessions/bulk, v4 mirroring by @nkabardin in https://github.com/langfuse/langfuse/pull/14104
- feat: detect clickhouse versions on start and set compat flags automagically. by @sumerman in https://github.com/langfuse/langfuse/pull/14187
- feat(agent): Add conversation starters by @bezbac in https://github.com/langfuse/langfuse/pull/14188
- feat(monitors): bucket live preview by evaluation window by @marksalpeter in https://github.com/langfuse/langfuse/pull/14189
- feat(batch-export): read sessions from events table for v4 users by @niklassemmler in https://github.com/langfuse/langfuse/pull/14040
- feat(monitors): retire feature flag, raise org limit to 20 by @marksalpeter in https://github.com/langfuse/langfuse/pull/14199
- feat(support): update support severity drop down by @jannikmaierhoefer in https://github.com/langfuse/langfuse/pull/14190
- feat(evals): change icon for llm-as-a-judge evaluator set up by @marliessophie in https://github.com/langfuse/langfuse/pull/14207
- feat(corrections): remove beta label by @marliessophie in https://github.com/langfuse/langfuse/pull/14210
- feat(blob-exporter): gate isEnrichedBlobExportAvailable on V4 env vars for self-hosted by @niklassemmler in https://github.com/langfuse/langfuse/pull/14098
- feat(observations): checkboxes in traces/observations are more performant and support shift-selection by @nkabardin in https://github.com/langfuse/langfuse/pull/14021
Fixes
- fix(media): drop media link foreign keys by @hassiebp in https://github.com/langfuse/langfuse/pull/14170
- fix(worker): throttle Mixpanel integration exports by @nyarram in https://github.com/langfuse/langfuse/pull/13958
- fix(ui): resolve app theme in CodeBlock for dark mode syntax highlighting by @jannikmaierhoefer in https://github.com/langfuse/langfuse/pull/14200
- fix: gate sessions events session_id pushdown on 'any of' operator by @niklassemmler in https://github.com/langfuse/langfuse/pull/14202
- fix(batch-actions): snapshot v4 events-table flag for session batch actions by @niklassemmler in https://github.com/langfuse/langfuse/pull/14201
- fix(seeder): align agentic seed summaries by @nkabardin in https://github.com/langfuse/langfuse/pull/14209
- fix(batch-export): read scores trace metadata from events table for v4 users by @niklassemmler in https://github.com/langfuse/langfuse/pull/14203
- fix: refine API error logging and prompt validation handling by @sumerman in https://github.com/langfuse/langfuse/pull/14215
Chores
- refactor(public-api): redesign repository interfaces to accept pre-built FilterList by @niklassemmler in https://github.com/langfuse/langfuse/pull/14182
- chore(agent): Add feedback link to feature preview modal by @bezbac in https://github.com/langfuse/langfuse/pull/14197
- chore: decrease default LANGFUSE_MIXPANEL_FLUSH_DELAY_MS by @sumerman in https://github.com/langfuse/langfuse/pull/14198
- docs(agents): warn against is_deleted filter on ClickHouse reads by @niklassemmler in https://github.com/langfuse/langfuse/pull/13977
- docs(agents): update weekly production review skill by @maxdeichmann in https://github.com/langfuse/langfuse/pull/14208
- ci(deps): bump the github-actions group with 4 updates by @dependabot[bot] in https://github.com/langfuse/langfuse/pull/14212
New Contributors
- @nyarram made their first contribution in https://github.com/langfuse/langfuse/pull/13958
Full Changelog: https://github.com/langfuse/langfuse/compare/v3.184.1...v3.185.0
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
About langfuse
🪢 Open source LLM engineering platform: LLM Observability, metrics, evals, prompt management, playground, datasets. Integrates with OpenTelemetry, Langchain, OpenAI SDK, LiteLLM, and more. YC W23
Related context
Related tools
Beta — feedback welcome: [email protected]