What does Redeep do for you?

Redeep gives me one place to assign, follow, and review work done by AI. Each task lives on a visual board, so I can see what is waiting, what is underway, and what needs my attention — no trail of chat tabs.

You can ask one worker to research a market, another to turn a spreadsheet into a decision-ready report, a third to update a website. They run at the same time. You stay at the level of goals and outcomes, then inspect the work that comes back.

The simple version: write a task on a card, move it into action, and let Redeep carry the context and the work until the result is ready.

How is Redeep different from a chatbot?

A chatbot mainly answers in a conversation; a Redeep AI worker is a persistent work session that acts on the materials and tools behind the conversation. Give it a clear brief and it opens the files in a project, makes changes, runs calculations, uses a signed-in browser, and operates desktop apps with your approval — while you focus on something else.

The work persists. Close the app and come back later — the task, conversation, files, and progress are still there. When several tasks belong to one project, the board keeps the whole effort legible.

Best of all, parallel work doesn't become parallel chaos. Each worker gets an isolated copy of the project while it works, so two tasks can change the same project without overwriting one another. You review the changes before they join the main version.

See how this works in practice in our guide to running projects with a Kanban board.

Do you control the model, cost, and data?

Yes. Redeep runs on your Mac, uses provider accounts you choose, and stores project data locally. Bring an API key from OpenAI, Anthropic, Google, OpenRouter, or Vercel AI Gateway — you decide which provider handles the work and pay that provider directly.

For smaller jobs, you can run an on-device model on Apple Silicon. That route needs no provider key and sends nothing to a cloud model. I mix the two: local models for routine helper work, powerful cloud models for assignments that need deeper judgment.

Redeep is a native macOS app — not a website living in another browser tab, not a repackaged web app. It is free to download, and its local-first design keeps your files, history, and credentials on your machine. Keys live in the macOS Keychain.

What kind of work can Redeep handle?

Redeep comes with more than 30 built-in capabilities for everyday knowledge work and software projects. You don't have to wire together a pile of utilities before the first useful task.

  • Research: search the web, read pages, compare sources, and work inside a browser session that already has your logins.
  • Documents and data: read and write common files, query spreadsheets and databases, run Python analysis, and produce charts or shareable reports.
  • Software and websites: inspect a project, edit files, run tests, review changes, and manage version history.
  • Your Mac: with explicit permission, read the screen, click controls, type, and work in desktop apps.
  • Images and media: inspect images and generate new visual assets when a task calls for them.

Every action has a safety level. Reading can happen automatically; writing or running something can require approval; high-impact desktop actions always ask. You stay in control without supervising every harmless step.

The complete list is in the built-in tools documentation.

What is the easiest way to start?

Start with a task that has a clear finish line and materials already on your Mac. “Compare these three vendor proposals and write a one-page recommendation” beats “help with procurement.” “Review this project and fix the failing tests” beats “improve the code.”

Download Redeep, connect the provider you already use, and point a new task at the right folder. Describe the outcome, any boundaries, and what “done” looks like. The worker plans the route; you provide the destination.

You. Your keys. Your files. A team that can keep working while you move on.