LangChain
LangChain is a framework that eases developing with LLMs.
Vocabulary
- Chain
- Model
- Prompt
- Output parser
- Prompt template - Similar to a Mustache template. Replace prompt variables with user values at run time.
- LCEL - LangChain Expression Language - Uses pipe syntax, i.e.,
| - Runnable protocol
- stream
- invoke
- batch
Features
- Force models to call a tool
- OpenTelemetry support for metrics/analytics
Alternatives
- AutoGPT
- GPT-Engineer
- LlamaIndex
- Microsoft AutoGen
Contact
- Slack
- Email: support@langchain.dev
Resources
- https://chat.langchain.com/?threadId=8b948dfc-7b41-4ce0-8fbd-7dc387aa5fd0
- https://docs.smith.langchain.com/evaluation/how_to_guides/attach_user_feedback
- Lucas Soares taught Getting started with LangChain on O'Reilly
Traces
Testing
Evals
- openevals: https://github.com/langchain-ai/openevals
- agentevals: https://github.com/langchain-ai/agentevals
LangChain
Interactive graph
On this page
Vocabulary
Features
Alternatives
Contact
Resources
Traces
Testing
Evals