I’m trying to figure out if I can run unit tests using just the langsmith library without needing the paid web service. I found some documentation about unit testing with langsmith but I’m confused about what parts require payment. Can I use the library standalone for testing my code and see results locally? Or do I have to use their web platform to get any meaningful test results? I want to know if the web app is required or if I can just install the library and run tests without any external services. Has anyone used langsmith for testing without the web interface? Thanks for any help!
Works perfectly offline! I’ve been running it locally for months without paying anything. Just don’t set up API keys or it’ll try connecting to their servers. Yeah, their docs are confusing - they mix local and cloud stuff together.
Yeah, you can definitely use langsmith locally without paying for anything. The library’s free and runs independently.
I’ve used it for 6 months now. Just pip install it and start testing. All core functionality runs on your machine - no web calls needed.
You can create test cases, run evaluations, and get results locally. The library handles execution and dumps data right to your terminal.
The web platform’s for dashboards, team stuff, and cloud storage. For basic unit testing? The library’s plenty.
Just don’t accidentally configure it to send data to their servers. Double-check your environment variables and config.
This video explains what’s local vs cloud pretty well.
Yep, langsmith works fine without paying anything or using their web platform. The library’s open source and runs all tests locally. I’ve used it in CI/CD pipelines for months without touching their cloud stuff. Here’s the thing - langsmith has two parts: the free Python library and their paid platform. For unit tests, you just need the library. It’s got evaluation frameworks, test runners, and generates results locally. The web interface is only for pretty charts and team features. Don’t set LANGSMITH_API_KEY in your environment and everything runs offline. Actually faster too since there’s no network calls.
The langsmith library is completely free for local testing. I’ve been running it offline for months without any subscription. You can run evaluations, compare model outputs, and generate test reports right on your machine. The confusion comes from their docs mixing library features with platform features. Install langsmith via pip and you get the full testing framework that works independently. The web platform is optional - it just adds collaboration tools, persistent storage, and analytics dashboards. For solo development and unit testing, the library has everything you need including detailed logging and result analysis. Just don’t set any API keys in your environment if you want to stay completely local.