langfuse/langfuse: 🪢 Open source LLM engineering platform: LLM Observability, metrics, evals, prompt management, playground, datasets. Integrated with LlamaIndex, Langchain, OpenAI SDK, LiteLLM, etc. 🍊YC W23
Langfuse is used Github Discussions for Support and Feature Requests.
We are hiring. Join us in Product Engineering and Developer Relations.
- LLM Analytics: Track metrics (cost, latency, quality) and gain insights from dashboards and data exports (Analytics)
- LLM Evaluation: Collect and calculate marks for your LLM completion (Scores and Evaluations)
- Experiments: Track and test app behavior before deploying a new version
Managed deployment by the Langfuse team, several free-tiers (fun plans), no credit card required.
# Clone repository
git clone https://github.com/langfuse/langfuse.git
cd langfuse
# Run server and database
docker compose up -d
→ Learn more about deploying locally
Langfuse is simply self-hosted and continuously updated. It currently only requires a docker container and a postgres database.
→ Self-Hosting Instructions
Templated deployments: Railway, GCP, AWS, Azure, Kubernetes and more
You need a Langfuse public and secret key to get started. Sign up HERE and find them in your project settings.
Note: We recommend using our fully async, typed SDKs which allows you to instrument any LLM application with any underlying model. They are in the Python (Decorators) & JS/TS. SDKs are always the most fully featured and robust way to get data into Langfuse.
See the → Quick start to integrate Langfuse.
| UNITY | supports | Description |
|---|---|---|
| SDK | Python, JS/TS | Manual instrumentation using SDKs for complete flexibility. |
| OpenAI | Python, JS/TS | Automated instrumentation using a drop-in replacement for the OpenAI SDK. |
| Lanchain | Python, JS/TS | Automated instrumentation by passing the callback handler to the Langchain application. |
| LlamaIndex | Python | Automated instrumentation through the LlamaIndex callback system. |
| Haystack | Python | Automated instrumentation through the Haystack content tracing system. |
| LiteLLM | Python, JS/TS (proxy only) | Use any LLM as a drop in replacement for GPT. Use Azure, OpenAI, Cohere, Anthropic, Ollama, VLLM, Sagemaker, HuggingFace, Replicate (100+ LLMs). |
| Vercel AI SDK | JS/TS | TypeScript toolkit designed to help developers create AI-powered applications using React, Next.js, Vue, Svelte, Node.js. |
| maltreatment | Call the public API directly. OpenAPI spec available. |
Packages included with Langfuse:
| name | Description |
|---|---|
| Instructor | Library to get structured LLM outputs (JSON, Pydantic) |
| Diff | Open source LLM app development platform without code builder. |
| To be | Easily run open source LLMs on your own machine. |
| Mirascope | Python toolkit for building LLM applications. |
| Pouring | JS/TS no-code builder for customized LLM flows. |
| Langflow | The Python-based UI for LangChain, designed with react-flow to provide an effortless way to experiment and prototype flows. |
In order of preference the best way to communicate with us:
This repository is licensed to MIT, except for ee folders. See LICENSE and docs for more details.
GET routes to use data in downstream applications (eg embedded analytics). You can also access it easily through SDKs (docs).
We take data security and privacy very seriously. Please see our Security and Privacy page for more information.
By default, Langfuse automatically reports basic usage statistics of self-hosted instances to a centralized server (PostHog).
This helps us to:
- Understand how to use Langfuse and develop the most relevant features.
- Track overall usage for internal and external (eg fundraising) reporting.
None of the data is shared with third parties and does not include any sensitive information. We want to be super transparent about this and you can find the exact data we collect HERE.
You can choose by setting TELEMETRY_ENABLED=false.
Top open-source Python projects using Langfuse, ranked by stars (SOURCE):
https://opengraph.githubassets.com/19497c27b2e6fdd86f837e66d92ce2374f9cb922a9fd9f74ca12f42c6a3d915f/langfuse/langfuse
2024-12-17 13:43:29
