Install ProPR, or skip the server
Run npm install -g propr-cli and propr setup on a Linux Docker host (quickstart). ProPR Hosted runs a dedicated instance for you when you would rather not run one.
ProPR: the self-hosted Devin alternative for any agent
Devin, from Cognition, is a hosted autonomous engineer with its own workspace, plan step, and review. ProPR does that job on your own server with the agent you choose, from approved plan to pull request to comment-driven commits, on the subscriptions you already pay for.
vs
Devin
Worker or coordinator
Devin is the worker and workspace. ProPR coordinates different workers inside your own engineering process.
Devin pathSlack, Linear, Jira, or
/devin → hosted workspace → pull requestCognition hosts the compute, the models, and the session history.ProPR pathPlan → isolated workspace → GitHub PR → follow-up commitPick the agent per task; task history and logs stay on your server, and review happens in GitHub.
Switch from Devin to ProPR.
What you keep: your GitHub repositories and review process, and the Claude or ChatGPT plan your developers already hold, which ProPR runs through Claude Code or Codex.
The same ideas, ProPR's names
| Devin | ProPR |
|---|---|
| Session | Task |
| Devin's editable plan, optional approval wait | Plan, approved before code runs |
/devin comment on a pull request | Any plain PR comment |
/devin review and Auto-Fix | /review and /fix |
| Skills, Playbooks, and Knowledge | Agent instructions in .propr/workflow.yml |
| Cognition's model catalog | Agent and model label per task |
| app.devin.ai, Slack, Linear, Jira | GitHub issue, web UI, desktop app, CLI, MCP |
From zero to the first pull request
Add the same repository and sign in your agent
Open the web UI, sign in with GitHub, and add the repository. Sign in Claude Code with your Claude plan, or Codex with your ChatGPT plan; Devin took neither.
Start the first task and approve the plan
Take the issue you would have handed to Devin, add the AI label, and approve the plan ProPR posts. Several model labels on one issue produce one pull request per model.
Review the pull request in GitHub
Review the diff the way you review any pull request. A plain comment on it becomes the next commit on the same branch; /review posts scored findings and /fix applies them. The follow-up playbook walks one comment to its commit.
What stays with Devin: ProPR has no Slack, Linear, or Jira entry point and no live browser-and-IDE workspace to watch; tasks start from GitHub, the web UI, the desktop app, the CLI, or MCP.
Devin takes the whole task off your hands. It starts from app.devin.ai, Slack, Linear, Jira, the API, or a /devin comment on a GitHub pull request, researches the codebase, writes a plan you can edit or hold for approval, implements in a hosted workspace with shell, IDE, and browser, and opens the pull request. It replies to comments on PRs its sessions are working on, and Devin Review posts findings with an Auto-Fix session to address them. You pick the model from Cognition's catalog; Devin takes no third-party API keys. Plans run from free to $200 per month for individuals, with Teams at $80 plus $40 per seat, and Enterprise single-tenant VPC deployments that Cognition hosts.
ProPR centers on the pull request: each phase uses the agent that fits, every change starts from an approved plan, and people finish it in GitHub. Plain comments, with no command prefix, become follow-up commits from Claude Code, Codex, Antigravity CLI, Mistral Vibe, or OpenCode on the subscriptions or keys you already hold.
The source, the credentials, and the task history are yours.
Devin fits best when...
- You want to delegate whole tasks to one hosted autonomous engineer from Slack, Linear, Jira, or GitHub.
- A conversational workspace with built-in shell, IDE, and browser tools is how you want to supervise it.
- Cognition hosting the compute, models, and session history is acceptable.
ProPR fits best when...
- AI development runs through a planned PR path on your own server.
- You pick from Claude Code, Codex, Antigravity CLI, Mistral Vibe, OpenCode, subscriptions, or direct API access.
- Plain PR comments drive follow-up commits, and the workspace, logs, and source are yours.
A hosted autonomous teammate versus your own PR path.
| Question | Devin | ProPR |
|---|---|---|
| Hosted autonomous teammate is the product | Yes | No, it runs your agents |
| Plan before implementation | Editable plan, optional approval wait | Reviewed plan on every task |
| Opens the pull request | Yes | Yes |
| GitHub comments become follow-up commits | /devin on an open PR; replies on its own PRs | Any plain comment, any agent |
| PR review from a comment | /devin review, with Auto-Fix | /review scores, /fix applies |
| Choose Claude Code, Codex, Antigravity CLI, Mistral Vibe, OpenCode, or direct APIs per phase | No, Devin only; no third-party keys | Yes |
| Self-hosted task records and logs | No; Enterprise single-tenant VPC hosted by Cognition | Yes |
| Platform pricing beyond model costs | $0–$200/mo; Teams $80 plus $40 per seat | Free, Apache 2.0 |
Verified November 2026: Devin GitHub integration, Devin Review, Devin models, Devin enterprise deployment, Devin 2.0, devin.ai/pricing, Devin Playbooks, Devin Knowledge.