Repolode
Submit
pezzolabs

pezzolabs/pezzo

Open-source, cloud-native LLMOps platform for managing prompts, observing LLM calls, caching, and delivering prompt changes.

Open-source alternative to:Langfuse

View on GitHubWebsite
Stars
3,279
Forks
280
Open issues
56
Last commit
3 months ago

What is it?

Pezzo is an open-source LLMOps platform for teams that run prompts against large language models in production. It puts prompt design, version management, and collaboration in one console, so prompt changes can be delivered to applications without redeploying code.

The platform also covers operations. It records observability data for LLM calls to help troubleshoot issues, and offers caching to reduce cost and latency. The stack is built on PostgreSQL, ClickHouse, Redis, and SuperTokens, and it can be run locally with Docker Compose.

Official client libraries are available for Node.js and Python, with LangChain support listed as well. Each client provides prompt management, observability, and caching.

Key features

  • 01Centralized prompt design, versioning, and collaboration in one console
  • 02Instant delivery of prompt changes to applications
  • 03Observability and monitoring of LLM operations for troubleshooting
  • 04Caching to reduce LLM costs and latency
  • 05Node.js and Python clients, with LangChain support

Use cases

  • Managing and versioning prompts across several AI features in a product team
  • Monitoring LLM calls in production to find failures and costly requests
  • Updating prompts for a running application without a code deployment

Quick start

  1. 1

    Install npm dependencies

    npm install
  2. 2

    Start the infrastructure dependencies with Docker Compose

    docker-compose -f docker-compose.infra.yaml up
  3. 3

    Deploy Prisma migrations

    npx dotenv-cli -e apps/server/.env -- npx prisma migrate deploy --schema apps/server/prisma/schema.prisma
  4. 4

    Run the server and the Pezzo Console

    npx nx serve server

Things to know

  • Development setup requires Node.js 18+, Docker, and several infrastructure services such as PostgreSQL, ClickHouse, Redis and Supertokens.
  • The README's simplest path is the Docker Compose full-stack guide, which lives in external documentation rather than the README.

In collections

OpenTelemetry & observability platforms

Topics

#llmops#prompt-management#observability#llm#prompt-engineering#typescript#ai#devtools#gpt-3#gpt-4

FAQ

What is pezzo?
Open-source, cloud-native LLMOps platform for managing prompts, observing LLM calls, caching, and delivering prompt changes. Pezzo is an open-source LLMOps platform for teams that run prompts against large language models in production. It puts prompt design, version management, and collaboration in one console, so prompt changes can be delivered to applications without redeploying code.
Is pezzo free to use? What license does it have?
pezzo is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is pezzo still maintained?
Updated within 6 months. The last commit to pezzo was on Aug 21, 2026.
What can I use pezzo for?
Managing and versioning prompts across several AI features in a product team Monitoring LLM calls in production to find failures and costly requests Updating prompts for a running application without a code deployment
How do I install pezzo?
Install npm dependencies: npm install Start the infrastructure dependencies with Docker Compose: docker-compose -f docker-compose.infra.yaml up Deploy Prisma migrations: npx dotenv-cli -e apps/server/.env -- npx prisma migrate deploy --schema apps/server/prisma/schema.prisma Run the server and the Pezzo Console: npx nx serve server
Where can I get pezzo?
The source code and installation instructions are on GitHub: https://github.com/pezzolabs/pezzo

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects