Open Source Prompt Engineering Platform - Latitude

Latitude is an open-source platform designed for building and operating AI features in production. It emphasizes observability and evaluations to enhance the reliability of prompts used in large language models (LLMs).

Key features include:

• Observability - Capture prompts, inputs/outputs, tool calls, and latency/token usage from real traffic.
• Prompt Playground - Reproduce runs and iterate with real inputs while managing version changes.
• Datasets - Curate real examples for batch testing and regression suites.
• Evaluations - Built-in evaluations with LLM-as-judge and human-in-the-loop scoring.
• Experiments - Compare models/providers and prompt versions with measurable results.
• Reliability Loop - Turn production failures into repeatable fixes through annotations and automatic evaluations.

Latitude can be deployed as a managed cloud product or self-hosted, catering to various user needs. It supports most model providers and frameworks, making it a versatile choice for teams looking to enhance their AI capabilities.