Langfuse v4.43.0
Langfuse v4.43.0 adds role filtering for members, score creation directly in the annotation sidebar, and routing for TypeSafe decision-model evaluations via OpenRouter or Vercel AI Gateway, alongside an important SCIM organization-scoping fix.
原始内容为英文；当前页面提供中文导航与来源说明，具体事实请以原文为准。
为什么值得关注
Langfuse version 4.43.0 introduces several new features, fixes, and refactors across the platform: - Features : - Filter project and organization members by role. - Create scores directly from the annotation sidebar. - Introduce an internal view mode and associated components behind feature flags. - Support routing TypeSafe decision-model connections through OpenRouter or Vercel AI Gateway for evals. - Fixes & Performance : - SCIM: Properly scope Users/{id} reads and mutations to the caller's organization. - Full-Text Search (FTS): Optimize token prefilter decisions at build time rather than using an OR clause. - Web UI: Various UI improvements, including observation attributes raw/formatted toggles, tree metric heatmap adjustments, trace page titles defaulting to trace IDs, and standardized badge and typography styles. - Chores & Refactoring : - Migrate API key tables, verified domain settings, and legacy dropdowns to standardized design-system components. - Update assistant integration to opus 5.5.
可执行摘要
Langfuse release v4.43.0 updates evals to support routing TypeSafe decision-model connections through OpenRouter or Vercel AI Gateway, adds role filtering for project/org members, allows creating scores via the annotation sidebar, addresses an authorization scoping issue in SCIM Users/{id} endpoints, and includes various UI/FTS optimizations.
- Agent 实用度
- 82/100
- 可信度
- 96%
- 机器格式
- JSON + Markdown
开发者应核对什么
- Verify SCIM integrations function properly with the new organization-level scoping on Users/{id} endpoints.
- Optionally configure OpenRouter or Vercel AI Gateway if routing TypeSafe decision-model evaluation traffic through them.