Hypotize

FAQ

What is an agentic Growth OS?

The full stack a growth hacker works with, built for your AI agent. It includes a framework with the method, an MCP server with tools to read and update records, three skills that teach the agent the method, a plugin that installs it in Claude Code and Codex, and a web app your team works in. Unlike spreadsheet trackers, Hypotize enforces rules that the agent cannot break.

Which agents work with Hypotize?

Claude, ChatGPT, Grok Bot, Cursor, Codex, and any client that connects to a remote MCP server with OAuth. In Claude Code and Codex you can also install the plugin, which adds the three skills to the connection.

Does Hypotize work for CRO teams?

Yes. CRO teams use it for the work around each test: a backlog of test ideas, scoring with ICE or PIE, the stage of every test and an archive of results and learnings. Your A/B testing tool still runs the test and splits the traffic. Hypotize keeps what you tested, why you tested it and what it taught.

Do I need an AI agent to use Hypotize?

No, but that is where it pays off. The web app works on its own, with the same board, backlog and learnings. With an agent connected, the routine work moves to the agent: it finds what the team already tested, adds and scores ideas, moves tests through their stages and records outcomes. The app is where your team reviews that work.

What can a connected agent see and change?

Only what your own account and role allow. It cannot reach a project you cannot reach yourself, and every action goes through the same permission checks as the web app. Agents can vote in your framework, read scores, add ideas, move statuses, log interim readings as comments, and record outcomes.