MCP · Chat control

Plan, start, review and merge from chat.

Draft a plan, start a task or goal, read the diff and the review findings, apply the fixes you pick, and request the merge, from the chat client you already use. Every action lands in GitHub as a normal issue, commit or pull request.

Built into ProPR as an optional server. Set it up.

Compare the two connection routes →

MCP architecture

Your chat clientText requests and structured results

Direct/api/mcpYour instance OAuth

Through Connectmcp.propr.dev/mcpConnect OAuth → managed tunnel

Your ProPR instanceOne tool implementation for both routesPlans · goals · tasks · PR workflow

Both routes end at your instance's /api/mcp; chat receives the results you ask for.

Two routes. The same controls.

Both routes serve the same tools from your instance, limited by your permissions. Connect adds hosted sign-in and routing.

Direct MCP · Your instance

Connect to your own instance.

https://<instance>/api/mcp

Your chat client talks to your instance and signs in through its own OAuth flow. Connect never sees the traffic. Cloud chat hosts such as Claude and ChatGPT need a public HTTPS address for the instance.

Run your own ProPR instance →

Connect MCP · Managed tunnel · Plus

Sign in once, pick your installation.

https://mcp.propr.dev/mcp

The hosted gateway forwards calls through your installation's managed tunnel to the same endpoint. It needs an active Plus tunnel and an instance that trusts Connect.

Explore Connect MCP →

What chat can do.

You read findings before choosing fixes, and every change returns an ID you can come back to in a later conversation. Delegate from your chat assistant follows one task from the first request to the PR discussion.

Ask what needs you. Get two reviews on the PR.A chat client asks for the work overview, then requests a review with two models. GPT-6 Astra and Claude Fable 5.1 each post an independent review on public propr #2905, the same result a /review astra fable PR comment gives.

Find the work waiting on you

List the repositories, branches, and models your account can use. Name a repository, plan, goal, task, or PR loosely and ProPR resolves it. Ask for a briefing and ProPR puts blockers first, then running work, then what finished.

Example requestBrief me on ProPR. What is waiting on me, and which repository matches ‘payments’?

Turn intent into a published plan

Draft a plan, gather repository context, and refine the scope. Every change saves a revision you can restore. Publishing creates the GitHub issues; you choose which to implement and with which models.

Example requestDraft a retry plan for acme/payments. Keep the public API unchanged. Show me the revised plan before publishing issues.

Start work and steer it

Launch a one-off task, or start a goal and steer it: send direction, pause, resume, or cancel. Send follow-ups, cancel tasks, and run a saved automation.

Example requestStart the approved goal in acme/payments with the model I selected. Pause it before I change the acceptance criteria.

Inspect the evidence

Read progress, task events, changed files, diffs, visual previews, and artifacts. One overview call joins running and recent tasks with their PRs, review decisions, and checks.

Example requestShow the latest events, diff, and screenshots for the task linked to acme/payments#42. Explain the failure and link the relevant artifact.

Review and improve the PR

Request a review from one model or up to eight side by side, read the findings, and apply exactly the ones you pick. Run Ultrafix with a cycle limit, send a focused follow-up, or update the branch from its base.

Example requestReview acme/payments#57. Show the findings, then apply F2 and suggestion S1. Limit Ultrafix to two cycles and report each cycle's score.

Merge when you say so

Merging names the exact PR and head commit and needs your merge permission and green checks. Auto-merge is its own explicit choice.

Example requestInspect acme/payments#57 and its checks. Show the exact head commit and ask me before requesting the merge.

The MCP reference names every tool, resource and prompt.

Run ProPR from the Claude app.

Add your instance as a custom connector in Claude and the same connection serves claude.ai, Claude Desktop and the mobile apps. Ask what is waiting on you, approve a plan, and start a fix from your phone.

Catch up

Blockers first, then running work, then what finished.

AskWhat's waiting on me in ProPR?

Decide

Read the findings, pick the fixes, and let Ultrafix run.

AskReview the checkout PR with GPT-6 Astra and show me the findings. Apply the first two.

Start

Plan, approve, and launch with the model you choose.

AskPlan rate limiting for orders-api. Once I approve, start it with Claude Opus 5.5.

Every request leaves a trail you can follow.

Chat sessions end and connections drop. ProPR records what each request started and how it ended, so the next conversation starts from facts.

Receipts you can look up later

Every change returns an ID, and one call lists what you started in the last hour. A single overview joins running and recent tasks with their pull requests, reviews, checks and Ultrafix state.

Errors that say what to do next

Every failure names a stable code, the stage that failed, and whether a retry is safe. The documentation ships inside the release, so answers about settings match the instance you run.

Screenshots in the conversation

A task's captured screenshots come into chat as downscaled images. Videos come back as a GitHub link.

An access log for administrators

Logs → MCP Log records which tool each app called, the outcome, and the timing. Arguments and result bodies never enter the log.

The MCP Log: each tool call with its connected app, outcome and duration; one call is denied with FORBIDDEN_REPOSITORY — The log marks a call outside the grant as denied.

Authorize the app. Bound every action.

You sign in and consent in your browser, pick the repositories the app may touch, and leave optional scopes unchecked. Connected apps shows each app's permissions and last use, and revokes any of them in one click. An app never gets more access than your account has.

Non-secret configuration is part of the same conversation: add repositories and agents, set branches, models, labels and policies, and tune indexing, review and Ultrafix defaults. Logins, credentials, environment variables, custom images and wider grants stay in the browser.

The MCP consent screen’s repository picker: two of four Northwind repositories selected, the starred one first, with Select all, Clear and a filter box — The consent screen grants access repository by repository.

Choose the route with its data boundary in view.

Direct MCP runs between your chat host and your instance and bypasses Connect. The Connect gateway carries tool arguments and results, which can include prompts, source code, diffs, and artifacts, with no end-to-end encryption through it.

Repository clones, agents, execution, and task state stay self-hosted. Whatever the conversation returns reaches your chat host and its model service, under that host's policies.

Read the full trust boundary →

Availability and setup.

MCP ships with ProPR, switched off. An administrator turns it on in Settings → Integrations & MCP → MCP Server, copies the connection URL, and sets the scope ceiling no client can exceed.

Follow the chat assistant guide to connect a client, and the MCP reference for scopes, consent, connected apps, and the full tool list.

MCP server settings: enabled, the connection URL to paste into a chat client, the scopes an admin allows, and a link to connected apps — Administrators enable MCP, copy the connection URL, and set the scope ceiling in Settings.