# Kindgi

> Build agents and flows in your own code, run them durably, and keep every step on the record.

Kindgi runs AI agents and flows beside your application. You define tools (your own
TypeScript or Python code), agents and flows in a pack; the Kindgi runtime runs them
durably, calls your code over HTTP, and records every step in a journal. The SDKs
(`@kindgi/sdk` on npm, `kindgi` on PyPI) and the CLI (`@kindgi/cli`) are Apache-2.0.
TypeScript projects run the CLI as `pnpm exec kindgi`; Python projects as
`npx --yes @kindgi/cli@<minor>`.

## Documentation Sets

- [Abridged documentation](https://docs.kindgi.com/llms-small.txt): a compact version of the documentation for Kindgi, with non-essential content removed
- [Complete documentation](https://docs.kindgi.com/llms-full.txt): the full documentation for Kindgi
- [Start](https://docs.kindgi.com/_llms-txt/start.txt): install, quickstarts, existing apps
- [Tutorials](https://docs.kindgi.com/_llms-txt/tutorials.txt): build something, step by step
- [Guides](https://docs.kindgi.com/_llms-txt/guides.txt): how to do one thing
- [Concepts](https://docs.kindgi.com/_llms-txt/concepts.txt)
- [Deploy](https://docs.kindgi.com/_llms-txt/deploy.txt)
- [CLI reference](https://docs.kindgi.com/_llms-txt/cli-reference.txt)
- [Python SDK reference](https://docs.kindgi.com/_llms-txt/python-sdk-reference.txt)

## Notes

- The complete documentation includes all content from the official documentation
- The content is automatically generated from the same source as the official documentation