Welcome!
Open Source text analytics platform for LLM apps
phospho is an open source text analytics platform to finally get your LLM app ready for production.
Monitor interactions between your LLM app and your users. Test changes before pushing to prod. Get real time AI analytics and product-level insights to improve your LLM app.
Keywords: logging, automatic evaluations, experiments, A/B tests, user feedback, testing
Get Started
Get started now
Setting up phospho is easy, and you can get started in minutes.
Don’t have an LLM app? Try phospho in Google Colab
Overview
How does it work?
Log
You log the tasks completed by your LLM app (ex: what the user asked, what the assistant answered)
Analyze
We run custom evaluation pipelines on your transcripts. No OpenAI key needed. (ex: detect when the assistant couldn’t answer the user’s question)
Explore
You explore your LLM app evaluations on the phospho dashboard and with integrations.
Key features
Logging
Log every interaction with your LLM app in a non-invasive way. No monkey patch, no proxy, no framework.
Automatic event detection
Define events relevant to you, phospho detect them and can warn you when they happen. Take action.
Automatic Evaluation
phospho classify interactions as successes or failures according to your own definition.
User Feedback
Collect and attach user feedbacks to specific interactions. All in one place.
Review and Label
Let your team review and annotate flagged interactions. Collaborate with non-technical team members.
Offline Testing
Run tests and get success scores to improve your app before releasing it
Eager to see it in action? Get started in minutes.
Was this page helpful?