Skip to content

langfuse

v3.193.0 Feature

This release adds 5 notable features for engineering teams evaluating rollout.

Published 1mo Tracing
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

analytics autogen evaluation langchain large-language-models llama-index
+12 more
llm llm-evaluation llm-observability llmops monitoring observability openai playground prompt-engineering prompt-management self-hosted ycombinator

Summary

AI summary

Broad release touches Chores, deps, web, and agent.

Full changelog

What's Changed

Features

  • feat(search-bar): support quoted score/metadata names with spaces — @nkabardin (#14367)
  • feat(web): inline edit/delete action icons for datasets, evaluators, and annotation queues — @wochinge (#14335)
  • feat(agent): remove in-app agent feature flag — @bezbac (#14364)
  • feat(ui): allow all users to book a call — @annabellscha (#14391)
  • feat(slack): add Slack marketplace install flow — @wochinge (#14350)
  • feat(worker): count started/incomplete blob export attempts (LFE-10407) — @niklassemmler (#14396)
  • feat(pricing): add default model price audit workflow — @hassiebp (#14289)
  • feat(blob-storage): log non-reversible S3 credential shape on signing failures — @niklassemmler (#14401)
  • feat(web): guide users to add other models in LLM connection form — @jannikmaierhoefer (#14405)

Fixes

  • fix(web): trace only 5xx tRPC errors — @maxdeichmann (#14385)
  • fix(agent): update system prompt to exclude langfuse-* envs — @bezbac (#14366)
  • fix(agent): minor styling fixes — @bezbac (#14368)
  • fix(annotation-queues): make queue dropdown scrollable — @xstraven (#14378)
  • fix(tracing): send batchIO I/O queries as POST to avoid HTTP 431 — @nkabardin (#14393)
  • fix(monitors): populate Level and Type filter options from API — @yurifedoseev (#14268)
  • fix(worker): resolve real ECS host id instead of os.hostname() 0.0.0.0 (LFE-10388) — @niklassemmler (#14386)
  • fix(evals): enhance evaluator onboarding — @marliessophie (#14394)
  • fix(slack): prepend basePath to OAuth callback redirects — @wochinge (#14397)
  • fix: turn public API errors caused by oversized response bodies into 4xx — @sumerman (#14398)
  • fix(llm): route Vertex Claude through Anthropic client — @wochinge (#14388)
  • perf: restrict scores subquery to relevant trace IDs in dataset runs CTEs — @marliessophie (#14259)
  • perf(observations-v2): use LEFT ANY JOIN, remove shadow query system (LFE-10410) — @niklassemmler (#14395)

Chores

  • chore(deps): degrade nodemailer back to v7 — @nimarb (#14376)
  • chore(deps): bump next to 16.2.9 — @nimarb (#14390)
  • chore(deps): bump turborepo to 2.9.18 — @nimarb (#14389)
  • chore(deps): bump release-it to 20.2.0 — @nimarb (#14362)
  • chore(deps): bump undici to 7.28 — @nimarb (#14404)
  • chore(ci): rename secret to CLAUDE_API_KEY as it already exists in repo — @hassiebp (#14406)
  • chore: revert CODEOWNERS for GitHub config — @wochinge (#14407)
  • docs: add Scarf web traffic pixel to README — @jannikmaierhoefer (#14392)

New Contributors

  • @xstraven made their first contribution in https://github.com/langfuse/langfuse/pull/14378
  • @yurifedoseev made their first contribution in https://github.com/langfuse/langfuse/pull/14268

Full Changelog: https://github.com/langfuse/langfuse/compare/v3.192.2...v3.193.0

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

Track langfuse

Get notified when new releases ship.

Sign up free

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

All releases →

Related context

Beta — feedback welcome: [email protected]