Pythe

What Pythe is

Pythe is an AI cofounder with a team of AI employees. It plans your startup from your own facts, then does the work on its own computers. You approve anything that leaves.

Updated 2026-10-13

Pythe does two jobs for a founder. First, it works out what you should do next. Second, it puts an AI team on the work.

The two halves

The plan. You tell Pythe about your company with a few sentences, a website, a document or a public GitHub repo. Pythe writes up what it understood and you correct it. Then it researches your market and builds a plan for your stage, split into tracks like Product, Growth and Revenue. See How the plan is built.

The team. Pytheus is your AI cofounder. It knows your company, your plan and your Library. It can answer, research and run code. When a job is bigger, it briefs an AI employee, or proposes a hire that you confirm with one click: a researcher, writer, engineer, analyst, outreach (SDR) or operator. Each employee has its own cloud computer, so it can browse, run code, make Excel, Word, PowerPoint and PDF files, and build websites.

What makes it safe to hand work over

  • Employees can read, research, draft and write on their own computers without asking.
  • Anything that leaves Pythe, such as an email, a Slack post, a Notion page, a CRM record or a pull request, waits for your approval. Publishing a website is never skipped by autopilot.
  • The rule is enforced in code, not in a prompt. See Guardrails.
  • Your own AI tools can use Pythe too, but they cannot approve anything. See The MCP server.

What Pythe is not

  • It is not a generic chatbot. Everything it says starts from your company profile, your plan and what your team has already done.
  • It does not post, send or publish on its own. The one exception is autopilot, which you turn on per employee and per connector on Premium, and which never covers publishing a website.
  • It does not invent customers, metrics or quotes. Employees are told to mark what they could not verify.

Who it is for

Solo founders and small teams, from an idea to early revenue and beyond. The plan covers seven stages: Idea, Validation, Build, Early Traction, Product-Market Fit, Growth and Scale. See The seven stages.

Where to go next

Related