amosjyng/vcr-langchain

Record and replay LLM interactions for langchain

35
/ 100
Emerging

This tool helps developers working with LangChain applications to record and replay interactions with large language models (LLMs) and other external tools. It takes your existing LangChain code and, after an initial run, stores the LLM responses and tool outputs. Subsequent runs then instantly provide these stored outputs, making your tests faster, deterministic, and free from external dependencies or side effects. This is primarily for software developers and quality assurance engineers.

No commits in the last 6 months.

Use this if you are a Python developer building applications with LangChain and want to make your unit tests for LLM logic faster, more reliable, and independent of external API calls.

Not ideal if you need to record and replay every possible LangChain functionality, as it doesn't currently hijack all features for recording.

software-testing developer-tooling LLM-development CI/CD-optimization
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 10 / 25

How are scores calculated?

Stars

81

Forks

6

Language

Python

License

MIT

Last pushed

Jun 26, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/amosjyng/vcr-langchain"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.