Pezzo
A self-hosted LLMOps platform for prompt versioning, monitoring and caching, with Node.js, Python and LangChain clients. Apache 2.0 licensed.
Pezzo is an open-source platform for developers and teams managing prompts and monitoring LLM applications. You can run the full stack locally with Docker Compose, keeping the prompt management and monitoring platform on infrastructure you control. Its source code uses the Apache 2.0 license.
Prompt design, version management and collaboration share one console. Teams can manage prompts centrally and deliver prompt changes to their applications, with monitoring and troubleshooting tools alongside them. This makes it relevant to projects where prompts need ongoing maintenance and several people contribute to AI behavior.
Pezzo supports Node.js, Python and LangChain clients. Each supports prompt management, observability and caching, so the choice of client doesn't restrict those core capabilities. Caching can reduce costs and latency, while monitoring helps teams investigate issues in their AI operations.
The platform has a server and a browser-accessible console. Its infrastructure uses PostgreSQL, ClickHouse, Redis and Supertokens, all open-source technologies that can run through Docker Compose. Pezzo also supports OpenAI's GPT-4o. Requests to OpenAI use that external model service even when Pezzo itself is self-hosted.
