Skip to content

langfuse

v3.203.0 Feature

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

Published 25d 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, web, trace, and tracing.

Full changelog

What's Changed

Features

  • feat(dashboard): chart polish, hover interactivity, and cross-chart sync (LFE-10549) by @nkabardin in https://github.com/langfuse/langfuse/pull/14618
  • feat(web): polish media tags in json viewers by @wochinge in https://github.com/langfuse/langfuse/pull/14630
  • feat(trace): show tool call arguments on hover by @hassiebp in https://github.com/langfuse/langfuse/pull/14565
  • feat(redis): add optional TLS for Redis Sentinel by @Devenir-Glorieux in https://github.com/langfuse/langfuse/pull/13360
  • feat(agents): support targeted model price audits by @hassiebp in https://github.com/langfuse/langfuse/pull/14673
  • feat(agent): Mutating MCP calls & HITL by @bezbac in https://github.com/langfuse/langfuse/pull/14399
  • feat: V4 historic backfill background migrations by @Steffen911 in https://github.com/langfuse/langfuse/pull/14206

Fixes

  • fix(assistant): better trace design by @nimarb in https://github.com/langfuse/langfuse/pull/14638
  • fix(v4): coarsen migration timeline buckets by @maxdeichmann in https://github.com/langfuse/langfuse/pull/14641
  • fix(openai-models): add gpt-5.5 and gpt-5.5-pro to model selector by @hassiebp in https://github.com/langfuse/langfuse/pull/14635
  • fix(otel): map GenAI IO on AI SDK spans by @hassiebp in https://github.com/langfuse/langfuse/pull/14655
  • fix(tracing): parse standalone Responses tool calls by @hassiebp in https://github.com/langfuse/langfuse/pull/14654
  • fix(datasets): show expected output on dataset item experiments page by @marliessophie in https://github.com/langfuse/langfuse/pull/14649
  • fix(shared): retry GCS signed-URL generation on transient signing failures by @vismaytiwari in https://github.com/langfuse/langfuse/pull/14637
  • fix(web): show fallback for failed media previews by @wochinge in https://github.com/langfuse/langfuse/pull/14662
  • fix(web): allow storage upload URLs in CSP by @wochinge in https://github.com/langfuse/langfuse/pull/14665
  • fix(trace): deduplicate observations by id so colliding ids can't crash trace load (LFE-10588) by @nkabardin in https://github.com/langfuse/langfuse/pull/14669
  • fix(tracing): avoid LangGraph for Responses tool calls by @hassiebp in https://github.com/langfuse/langfuse/pull/14671
  • fix(events): restore trace-level score filtering in v4 (LFE-10596) by @nkabardin in https://github.com/langfuse/langfuse/pull/14675
  • fix(trace): dedupe observations at the source so the detail panel matches the timeline (LFE-10588) by @nkabardin in https://github.com/langfuse/langfuse/pull/14678
  • fix(trace): decode unicode escapes in trace JSON download by @toyayuto in https://github.com/langfuse/langfuse/pull/14667
  • fix(experiments): thin x-axis ticks so long-name charts stop smearing (LFE-10583) by @nkabardin in https://github.com/langfuse/langfuse/pull/14670
  • fix(mcp): Enable filtering observations by metadata by @bezbac in https://github.com/langfuse/langfuse/pull/14584

Chores

  • chore(shared): tag dataset item write counts by @wochinge in https://github.com/langfuse/langfuse/pull/14639
  • refactor(web): Move LangfuseIcon to design-system folder by @bezbac in https://github.com/langfuse/langfuse/pull/14634
  • refactor(web): Move Switch to design-system folder by @bezbac in https://github.com/langfuse/langfuse/pull/14653
  • chore(pricing): update default model prices by @langfuse-bot in https://github.com/langfuse/langfuse/pull/14623
  • chore(worker): drop project_id tag from export volume metric for API integrations by @niklassemmler in https://github.com/langfuse/langfuse/pull/14656
  • chore(pricing): update default model prices by @langfuse-bot in https://github.com/langfuse/langfuse/pull/14664

New Contributors

  • @vismaytiwari made their first contribution in https://github.com/langfuse/langfuse/pull/14637
  • @Devenir-Glorieux made their first contribution in https://github.com/langfuse/langfuse/pull/13360

Full Changelog: https://github.com/langfuse/langfuse/compare/v3.202.1...v3.203.0

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]