This repository contains a set of principles for building Reliable Agentic AI. It originates in the Summer of 2025 with some of the key practitioners in the field, bridging cutting-edge OSS AI developments with established state-of-the-art distributed systems that power millions of devices already, as well as novel AI-first programming languages, LLM programming approaches, and industrial-strength agentic frameworks.
Read the Reliable Agentic AI Manifesto
See SIGNATORIES.md for the list of the original authors. If you agree with the manifesto and want to sign it and join the community, please create a pull request and add your name to the bottom of the SIGNATORIES.md file.
To provide your feedback and sign the Manifesto in the easiest way, please asnwer a short survey:
Answer the Reliable Agentic AI Survey #1
You'll be able to sign it by providing your name in the end. If you want to receive the updates, leave your email there, and share any additional principles of the Reliable Agentic AI that you think are important and missing from the Manifesto so far.