Langfuse Roadmap

Langfuse is open source and we want to be fully transparent what we’re working on and what’s next. This roadmap is a living document and we’ll update it as we make progress.

🚀 Released

10 most recent changelog items:

• Protected prompt labels(Apr 2, 2025)
• Playground support for Gemini 2.5 Pro Experimental(Mar 31, 2025)
• Tool Calling and Structured Output in Playground(Mar 28, 2025)
• Traces Table Peek View(Mar 21, 2025)
• New Prompt View(Mar 19, 2025)
• New Trace View(Mar 19, 2025)
• OpenAI Response API support in SDKs(Mar 17, 2025)
• Batch-export Scores via UI(Mar 13, 2025)
• Public API for Annotation Queues(Mar 13, 2025)
• Prompt Composability(Mar 12, 2025)

🚧 In progress

• Tracing
  • Full-text search: Search across the inputs/outputs of traces, sessions, and datasets (#939)
  • Analytics
    • Customizable dashboards: Create and share custom dashboards based on metrics extracted from traces (#1011)
    • Query API for custom metric aggregations to replace the static daily metrics API
  • Make Langfuse SDKs OpenTelemetry native, Langfuse Server already supports OpenTelemetry (docs)
• Evaluation
  • Session-level scores (#2728)
  • Improvements to core eval views (e.g. compare run view)
  • Annotate dataset experiments
• UI/UX: improvements of all core product features
• Admin/SCIM API to programmatically manage organizations, projects, and users (#1007)
• Self-host documentation: provide better guidance on selecting the right deployment options and scaling Langfuse

🔮 Planned

• Agent Observability: Improve native support for agentic applications
  • Generalized agent graphs (#2669), beta available for LangGraph (docs)
  • Filtering for tool calls used within an execution
  • Opinionated agent evaluations
• Evaluation
  • Simplified configuration of in-UI evals
  • More llm-as-a-judge templates for RAG, Agents, and conversational applications
  • Non-LLM evaluators: classifiers, custom code, regex, etc.
  • Evals on repeated spans/observations within a trace
  • Comparisons of different evaluation metrics
  • Improvements to Langfuse SDK to simplify creation of experiments
  • End-to-end examples on how to run Langfuse Experiments as regression tests in CI/CD pipelines
  • Add tracing to LLM-as-a-judge evals for better observability (debugging and cost tracking)
• UI
  • Filters: saved views, new filter UI on tables
• Prompt Management
  • Placeholders for chat messages (#2210)
  • Track prompt variables in production tracing, simplify adding to datasets and running prompt experiments
  • LLM-assisted prompt engineering
  • Folders (#4874)
  • A/B testing
  • Native support for tool calls (#2624)
• Data Platform
  • Webhooks to subscribe to changes within your Langfuse project (#1033)
  • Alerts: Create alerts for custom metric thresholds, errors, etc.
  • Rule-based routing between Langfuse product features
• Prompt Experiments & Playground
  • Split view to compare different LLMs and prompts side by side before running a prompt experiment
  • Add tool-call and structured output support to prompt experiments
• Langfuse Cloud
  • HIPAA compliance
  • Usage based billing alerts

🙏 Feature requests and bug reports

The best way to support Langfuse is to share your feedback, report bugs, and upvote on ideas suggested by others.

Feature requests

Bug reports