langfuse

v3.214.0 Feature

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

Published 11d Tracing

View tool

✓ 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

Affected surfaces

auth

Summary

AI summary

Broad release touches Chores, otel, mcp, and agent.

Changes in this release

Type	Severity	Summary	CVE
Feature
Feature	Low	Adds NEXTAUTH_COOKIE_NAME_SUFFIX to namespace auth cookies across instances. Adds NEXTAUTH_COOKIE_NAME_SUFFIX to namespace auth cookies across instances. Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Low	Adds hideXAxisLabels option to charts for improved readability. Adds hideXAxisLabels option to charts for improved readability. Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Low	Adds listMonitors and getMonitor tools to mcp and monitors. Adds listMonitors and getMonitor tools to mcp and monitors. Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Low	Adds metadata filter support in monitor queries. Adds metadata filter support in monitor queries. Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Low	Adds x-langfuse-ingestion-version header tagging on OTel API spans. Adds x-langfuse-ingestion-version header tagging on OTel API spans. Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Low	Adds default root obs filter on for app roots in filters. Adds default root obs filter on for app roots in filters. Source: llm_adapter@2026-07-15 Confidence: high	—
Feature	Low	Emits code eval dispatch duration distribution metric. Emits code eval dispatch duration distribution metric. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Feature	Low	Removes propagation filter warning popup in evals. Removes propagation filter warning popup in evals. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Feature	Low	Adds support-agent scenario – handcrafted demo‑grade agent trace to seeder. Adds support-agent scenario – handcrafted demo‑grade agent trace to seeder. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix
Bugfix	Medium	Fixes serialization of int64 fields as strings in OTel. Fixes serialization of int64 fields as strings in OTel. Source: llm_adapter@2026-07-15 Confidence: low	—
Bugfix	Medium	Fixes fallback when crypto.randomUUID is unavailable on non‑secure origins. Fixes fallback when crypto.randomUUID is unavailable on non‑secure origins. Source: llm_adapter@2026-07-15 Confidence: low	—
Bugfix	Medium	Fixes bound trace‑card I/O in session detail view for deep traces. Fixes bound trace‑card I/O in session detail view for deep traces. Source: llm_adapter@2026-07-15 Confidence: low	—
Bugfix	Medium	Fixes bound trace‑card I/O in session detail view for performance (LFE-10958). Fixes bound trace‑card I/O in session detail view for performance (LFE-10958). Source: llm_adapter@2026-07-15 Confidence: low	—
Bugfix	Low	Serializes int64 fields as strings in OTel instead of protobufjs Long internals. Serializes int64 fields as strings in OTel instead of protobufjs Long internals. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Adds gradient animation to loading state in agent. Adds gradient animation to loading state in agent. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Makes demo seeder import‑safe outside a monorepo checkout. Makes demo seeder import‑safe outside a monorepo checkout. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Normalizes pricing for Gemini reasoning tokens. Normalizes pricing for Gemini reasoning tokens. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Adds user agent to mcp spans. Adds user agent to mcp spans. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Streams core data S3 exports with keyset pagination in worker. Streams core data S3 exports with keyset pagination in worker. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Allows custom OpenAI provider options in llm. Allows custom OpenAI provider options in llm. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Makes user aware of screen context in agent. Makes user aware of screen context in agent. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Adds validation for experiment-item-version in otel. Adds validation for experiment-item-version in otel. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Aligns eval rule limit tests with 500 cap. Aligns eval rule limit tests with 500 cap. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Routes next‑auth logging through structured logger in auth. Routes next‑auth logging through structured logger in auth. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Resolves preview-comment action from base, builds from PR head. Resolves preview-comment action from base, builds from PR head. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Bugfix	Low	Keeps tree and timeline layouts readable on extremely deep traces (LFE-10959). Keeps tree and timeline layouts readable on extremely deep traces (LFE-10959). Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Refactor	Low	Makes AI SDK the sole completion runtime in llm. Makes AI SDK the sole completion runtime in llm. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—
Refactor	Low	Uses native AI SDK errors in llm. Uses native AI SDK errors in llm. Source: granite4.1:30b@2026-07-15-audit Confidence: high	—

Full changelog

What's Changed

Features

feat(evals): emit code eval dispatch duration distribution metric by @wochinge in https://github.com/langfuse/langfuse/pull/15071
feat(monitors): allow metadata filter in monitor queries by @marksalpeter in https://github.com/langfuse/langfuse/pull/15033
feat(otel): tag x-langfuse-ingestion-version header on OTel API spans by @maxdeichmann in https://github.com/langfuse/langfuse/pull/15076
feat(filters): default root obs filter on for app roots by @nimarb in https://github.com/langfuse/langfuse/pull/15066
feat(mcp,monitors): Add listMonitors and getMonitor tools by @bezbac in https://github.com/langfuse/langfuse/pull/15065
feat(evals): remove propagation filter warning popup by @marliessophie in https://github.com/langfuse/langfuse/pull/15101
feat(seeder): add support-agent scenario — handcrafted demo-grade agent trace by @nkabardin in https://github.com/langfuse/langfuse/pull/14985
feat(charts): add hideXAxisLabels option to improve readability by @marliessophie in https://github.com/langfuse/langfuse/pull/15027
feat(auth): add NEXTAUTH_COOKIE_NAME_SUFFIX to namespace auth cookies across instances by @maxdeichmann in https://github.com/langfuse/langfuse/pull/15094

Fixes

fix(otel): serialize int64 fields as strings, not protobufjs Long internals by @he-yufeng in https://github.com/langfuse/langfuse/pull/13342
fix(agent): Add gradient animation to loading state by @bezbac in https://github.com/langfuse/langfuse/pull/15064
fix(seeder): make the demo seeder import-safe outside a monorepo checkout by @maxdeichmann in https://github.com/langfuse/langfuse/pull/15070
fix(pricing): price normalized Gemini reasoning tokens by @hassiebp in https://github.com/langfuse/langfuse/pull/15082
fix(mcp): Add user agent to mcp spans by @bezbac in https://github.com/langfuse/langfuse/pull/15078
fix(worker): stream core data S3 exports with keyset pagination by @Steffen911 in https://github.com/langfuse/langfuse/pull/15083
fix(llm): allow custom OpenAI provider options by @hassiebp in https://github.com/langfuse/langfuse/pull/15085
fix(agent): Make user aware of screen context by @bezbac in https://github.com/langfuse/langfuse/pull/15079
fix(otel): add validation for experiment-item-version by @marliessophie in https://github.com/langfuse/langfuse/pull/15062
fix(tests): align eval rule limit tests with 500 cap by @LysanderKie in https://github.com/langfuse/langfuse/pull/15108
fix(web): fall back when crypto.randomUUID is unavailable on non-secure origins (LFE-10858) by @nkabardin in https://github.com/langfuse/langfuse/pull/15058
fix(mcp): Instrument more auth level fields by @bezbac in https://github.com/langfuse/langfuse/pull/15111
fix(codex): guard trap against unbound tmp_dir in ensure_migrate_binary by @marksalpeter in https://github.com/langfuse/langfuse/pull/15112
fix(auth): route next-auth logging through structured logger by @Steffen911 in https://github.com/langfuse/langfuse/pull/15096
fix(preview): resolve preview-comment action from base, build from PR head by @maxdeichmann in https://github.com/langfuse/langfuse/pull/15075
fix(sessions): bound trace-card I/O in the session detail view (LFE-10958) by @nkabardin in https://github.com/langfuse/langfuse/pull/15124
fix(trace): keep tree and timeline layouts readable on extremely deep traces (LFE-10959) by @nkabardin in https://github.com/langfuse/langfuse/pull/15123

Chores

chore: preview write access by @maxdeichmann in https://github.com/langfuse/langfuse/pull/15060
chore(agent): run model price audit daily by @hassiebp in https://github.com/langfuse/langfuse/pull/15067
refactor(llm): make AI SDK the sole completion runtime by @hassiebp in https://github.com/langfuse/langfuse/pull/14979
chore(worker): Enable knip dep checking & remove unused deps by @bezbac in https://github.com/langfuse/langfuse/pull/15028
chore(unstable-llm-api): increase max eval rule count to 500 by @hassiebp in https://github.com/langfuse/langfuse/pull/15103
refactor(llm): use native AI SDK errors by @hassiebp in https://github.com/langfuse/langfuse/pull/15091
docs(contributing): add CLA license check troubleshooting steps by @marksalpeter in https://github.com/langfuse/langfuse/pull/15090
chore(events): extract shared event row-selection planning by @sumerman in https://github.com/langfuse/langfuse/pull/14974
chore(preview): add web/worker log commands to the preview PR comment by @maxdeichmann in https://github.com/langfuse/langfuse/pull/15127

New Contributors

@he-yufeng made their first contribution in https://github.com/langfuse/langfuse/pull/13342

Full Changelog: https://github.com/langfuse/langfuse/compare/v3.213.0...v3.214.0

View diff on GitHub

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

Share on X Share on Bluesky

Track langfuse

Get notified when new releases ship.

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 →