# CatalEx

> CatalEx is the workbench for AI-native product teams. Product managers, designers, engineers, QA engineers and their AI agents build the same product in one place, and every decision is captured as a by-product of the work. CatalEx is coming soon; the waitlist is open at https://catalex.co.

## The problem

Every tool a product team uses stores the work. None of them store the reasoning. When a team is four people and forty agents, that reasoning is the company.

- **Dark decisions**: a PM and a developer change how a feature works on a five-minute call. The decision is real and binding, and it is written down nowhere. Days later QA tests against the original spec, files a bug, and the team loses a day arguing. Nobody was wrong; the decision had no home.
- **Agent amnesia**: AI agents working in the codebase have no idea why anything is the way it is, so they confidently refactor away intentional work, like the one odd retry loop that was protecting production.

## What CatalEx does

- **One workbench, native surfaces for every role**: specs for product managers, a design canvas for designers, code for engineers, test suites and runs for QA. Each is meant to be a place to make work, not a read-only view of it.
- **Decisions captured automatically**: agents listen to calls, threads, comments and code diffs and propose decisions. A person confirms in one click. Each decision records what was chosen, what was rejected, why, who decided, and every artifact it affects.
- **Everything linked**: specs, designs, code, tests and decisions are connected, and every artifact shows the decisions it embodies.
- **Drift detection**: when a change makes the spec, design, code or tests disagree, CatalEx raises it to the owner of the other end straight away, the way a compiler flags an error.
- **Agents that know why**: AI agents read the decisions behind the work and cite the decision that authorised each change they make.
- **Bring your tools**: CatalEx does not replace Figma, Notion, Confluence, Jira, Linear, GitHub or Slack. It imports work from them and owns the space between them.

## Who it's for

AI-native product teams: small groups of product managers, designers, engineers and QA engineers working alongside many AI agents, who need specs, designs, code and tests to stay in agreement as they ship fast.

## Status

CatalEx is in development and launching soon. Features described here reflect what is being built and may change before launch. Join the waitlist at https://catalex.co to hear first.

## Links

- [Website and waitlist](https://catalex.co): CatalEx homepage, launch countdown and waitlist
- [Blog](https://blog.catalex.co): Writing on AI agents, evals, agent memory and running AI in production
- [LinkedIn](https://www.linkedin.com/company/catalex-ai): CatalEx on LinkedIn
- [Instagram](https://www.instagram.com/talksbyaditi): Talks by the CatalEx founder
- [GitHub](https://github.com/catalex-ai): CatalEx on GitHub
- [YouTube](https://www.youtube.com/@Catalex-ai): CatalEx on YouTube

## Optional

- [Documentation](https://docs.catalex.co): Documentation for the CatalEx agent platform
- [Previous website](https://catalex.co/legacy/): Archive of the earlier CatalEx agent platform site
- [Privacy Policy](https://docs.catalex.co/privacy-policy): How CatalEx handles data
- [Terms of Service](https://docs.catalex.co/terms-of-service): Terms of use