pezzo

Pezzo is an open-source, cloud-native LLMOps platform aimed at teams building with large language models. According to its README, it is designed to streamline prompt design, version management, instant delivery, collaboration, troubleshooting and observability. Key capabilities described in the repository: - Prompt management: create, version and manage prompts in one place, with instant delivery of changes. - Observability: observe and monitor AI operations and troubleshoot issues. - Caching: the README states caching can help save on costs and latency. - Collaboration: manage prompts and AI changes collaboratively. Supported clients listed in the README include Node.js (@pezzo/client on npm), Python (on PyPI) and LangChain. The feature matrix shows prompt management, observability and caching supported across these clients. The project invites issues requesting additional clients. Architecture and deployment: Pezzo is described as fully cloud-native and relies on open-source technologies including PostgreSQL, ClickHouse, Redis and Supertokens. A Docker Compose setup is documented for running the full stack locally, and a development mode is described using Node.js 18+, Docker, Nx, Prisma migrations and GraphQL codegen. The server exposes a health endpoint, and the Pezzo Console runs as a separate web application. Documentation, tutorials and recipes for various use cases and LLM providers are available on the official docs site. The project welcomes community contributions via pull requests and issues, with a CONTRIBUTING.md and a Contributor Covenant code of conduct. Source code is released under the Apache 2.0 License.