Monitor AI errors with Latitude, for startup founders, backed by 4.4k+ GitHub stars
4,416 stars357 forksTypeScriptQuality 8/10Updated 7/12/2026100% free ยท open source
What it does
Latitude monitors AI errors and provides insights to improve model performance, specifically designed for startup founders to track and resolve AI issues efficiently.
โขWhen deploying AI models to production and wanting to monitor their performance in real-time
โขWhen experiencing unexplained errors or inconsistencies in AI model outputs
โขWhen needing to compare and contrast different AI model versions or configurations
Quick start
1Run `npm install` in the cloned repository to set up the project dependencies
2Configure the `latitude.config.ts` file to specify the AI model and monitoring settings
3Start the Latitude server using `npm run start`
4Access the Latitude web interface at `http://localhost:3000` to view AI model performance metrics
5Use the `latitudectl` command-line tool to manage and monitor AI models, such as `latitudectl models list` to list available models
Ready-to-paste prompt
Run `latitudectl models create --name my-model --type huggingface --endpoint https://my-model-endpoint.com` to create a new AI model configuration in Latitude
Heads up: Make sure to have Node.js (version 16 or higher) and npm installed on your system, as Latitude requires these dependencies to run
Saves to your device
Topics
agent-monitoring
agent-observability
ai-error-monitor
ai-monitoring
ai-observability
llm-error-tracing
llm-observability
llm-tools
What's inside โ free to inspect
No purchase needed
Read the entire source before you build โ unlike paid marketplaces that hide it behind a buy button.