Skip to main content
AI Choice Engine
← All setup guides

Cognition · Setup guide

How to set up Devin Cloud

Sign up for Devin Cloud in the browser, connect a repository with least privilege, and assign a small ticket-style task before any production repo access. This is not Devin Desktop (formerly Windsurf).

Devin Cloud

browser · checked October 2, 2026

~10 minBeginner

The “Open Devin Cloud” link goes to the vendor site. It is not an affiliate-tracked link on this page — no commission is claimed for that click. Affiliate disclosure.

Before you begin

  • A modern desktop browser.
  • A Cognition / Devin account with access to Devin Cloud.
  • A non-production repository or fork for the first trial.
  • Clear permission from whoever owns the GitHub/GitLab org before connecting.

Authentication

Browser account for Cognition / Devin Cloud

Account: Cognition account with Devin Cloud access

Cloud is distinct from Desktop, while eligible self-serve plans share quota and linked ChatGPT billing. Pro/Max/Teams can link per-user ChatGPT Plus/Pro for eligible GPT requests; other ensemble models still consume Devin quota and unavailable ChatGPT usage falls back automatically.

Install or open it

1. Get the harness ready

These steps are written for beginners. On Mac, Terminal is the black/white text window where you type commands. On Windows, use PowerShell or Terminal. Use the OS tabs to hide steps that are not for your computer, then copy one command at a time.

  1. 01

    Open Devin Cloud

    In Chrome, Safari, or Edge, go to https://app.devin.ai (marketing entry: https://devin.ai/cloud/) — not the Desktop IDE download. Create or sign into your Cognition account.

    Devin Cloud app
  2. 02

    Confirm you are on Devin Cloud

    Use the browser Cloud workspace. Devin Desktop is the separately installed IDE, whose primary bundled local agent is Devin Local; old Cascade histories and supported ACP agents do not establish Cloud access.

  3. 03

    Connect source control carefully

    Authorize the intended source-control integration with its required administrator, select the practice repository, and finish indexing/machine setup. Start in Ask for codebase exploration without actual-code edits. Repository access and session security profiles are separate; review both.

  4. 04

    Index the machine, then Ask before Agent

    Let Devin finish repo/machine setup and indexing. Start in Ask/plan mode for a read-only orientation before enabling Agent edits on a real ticket.

  5. 05

    Verify the cloud workspace

    Confirm the browser shows the Devin Cloud workspace and the practice repository before starting a session. If you are being asked to install an IDE, return to the previous product-name check.

    You should see: A Devin Cloud workspace with only the intended practice repository connected.

Repository or project

2. Start in the right place

Keep the first repository small, clean, and easy to roll back while you learn the harness’s permissions and context model.

  1. 01

    Create or fork a practice repo

    Use a disposable repository with tests and no production secrets. Note the default branch name.

  2. 02

    Open a new Devin Cloud session

    In the cloud workspace, start a new session bound to that repository only.

First run

Ticket-sized, reviewable work

Use this task to test the workflow before handing over a larger repository or production responsibility.

  1. 01Ask Devin Cloud to inventarise the repo and propose a one-file improvement with acceptance checks.
  2. 02Require a pull request or reviewable diff — do not merge automatically.
  3. 03Read every changed file before approving merge permissions.

Why this is safe: Cloud agents can open PRs and run remote compute; the first task should prove reviewability, not speed.

Stop and review if

  • The agent requests secrets, production database access, or org-admin scopes.
  • Changes land directly on main without a reviewable PR.
  • You cannot tell which machine or branch executed the work.

The operating boundary

Privacy, data, and cost

The setup is incomplete until you know what the harness can see, what account pays for it, and who approves a risky action.

Privacy

Devin Cloud runs outside your laptop. Cognition's current privacy policy permits content use for model training depending on service terms. Confirm the agreement and model/retention controls. Advertising opt-out is not model-training opt-out. Security profiles restrict configured network/MCP/git surfaces, but recommended profiles are overridable and changes reach existing sessions at their next wake/reboot.

Training on code: tier dependent

Confirm current Cognition enterprise and regional terms before regulated data.

Sensitive repos: Never connect production monorepos on day one; revoke SCM access after the pilot if you are not continuing.

Pricing

Pro $20/month, Max $200 and Teams $80 base plus $40/full seat use shared daily/weekly account allowance with extra usage at API rates. Enterprise terms differ; this is not a universal standalone pay-per-task Cloud tariff.

Model: hybrid

Watch out: Task size, model routing and retries affect quota. ChatGPT linking covers eligible GPT portions only and falls back to Devin billing when unavailable. A security profile's read-only git setting blocks push/PR creation; it does not make every local tool action read-only.

Estimate a seat stack

When setup goes wrong

  • You installed a desktop app by mistake.

    That is Devin Desktop. Use /ai-harnesses/setup/windsurf, or return to https://devin.ai/cloud/.

  • Repository connect fails or shows the wrong repos.

    Reconnect SCM with narrower permissions, confirm org SSO, and select the practice repo explicitly.

  • The session finishes with no reviewable PR.

    Tighten the prompt to require a PR, check branch permissions, and do not enable auto-merge for pilots.

Next steps

After the first successful run

Keep the pilot small, then compare surfaces and budget before a team rollout.

Setup details change quickly. The page was checked October 2, 2026; use the official source below before installing in a managed or production environment.