ProPR: the self-hosted alternative to Linear Coding Sessions

Linear Coding Sessions let the Linear Agent write code with Claude Code or Codex in Linear's cloud sandbox and draft the pull request. ProPR keeps planning, agent runs, pull requests, and follow-up commits on your server.

ProPR
vs
Linear

Tracker cloud or your server

Linear runs coding sessions inside its tracker and its cloud, billed from prepaid AI credits. ProPR runs on your infrastructure, keeps delivery in GitHub, and bills nothing itself.

Coding Session pathLinear issue → Linear cloud sandbox → draft pull requestReview continues in Linear's Reviews tab, on prepaid AI credits.
ProPR pathPlan → isolated workspace → GitHub PR → follow-up commitPlan, pull request, and follow-up commits in GitHub, on your server.

Switch from Coding Sessions to ProPR.

What you keep: your GitHub repositories and review process, Linear as the tracker, and the Claude or ChatGPT plan ProPR runs through Claude Code or Codex.

The same ideas, ProPR's names

Coding SessionsProPR
Coding SessionTask
Linear cloud sandboxIsolated workspace on your server
Draft pull request on the issuePull request on every task
Review handoff in the Reviews tabAny plain PR comment in GitHub
Prepaid AI creditsYour provider plan, no credits
Agent set per workspaceAgent chosen per task

From zero to the first pull request

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.

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 or Codex with the plan you already hold.

Start the first task and approve the plan

Add the AI label to the GitHub issue linked from the Linear issue, then approve the plan.

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 phone review playbook runs the loop from the GitHub mobile app.

What stays with Coding Sessions: ProPR has no Linear entry point; tasks start from GitHub issues, the web UI, the desktop app, the CLI, or MCP, and the pull request links back to the GitHub issue.

Assign an issue to the Linear Agent and it starts a session with the workspace's configured agent, Claude Code or Codex, in a Linear-managed cloud sandbox. It drafts the pull request, the diff appears on the issue, and review continues in Linear's Reviews tab, where you can hand review comments, rebases, and lint fixes back to the agent. Sessions are available on Linear's Free, Basic, Business, and Enterprise plans and draw from prepaid workspace AI credits: model tokens at provider rates plus sandbox time billed per 20-minute block.

ProPR starts with a plan, runs the agent in isolation on your infrastructure, opens the pull request, and turns GitHub comments into follow-up commits. Any supported agent runs the loop on the plan you already pay for, with no credits and no markup.

One run on Claude Opus 5.5 for PR #318: 19 minutes, 2.4M tokens in, 2.4% of the Claude session and 0.6% of the weekly window, an estimated $1.42
One run's record: duration, tokens, the share of the Claude session window it used, and the estimated cost, with no credits in between.

Linear Coding Sessions fits best when...

  • Issues already live in Linear and delegation should happen in the tracker.
  • Claude Code or Codex in Linear's cloud sandbox cover the work, and nobody wants to run a server.
  • Reviewing the diff inside Linear, then handing fixes back to the agent there, suits the team.
  • Prepaid AI credits are an acceptable way to pay for tokens and sandbox time.

ProPR fits best when...

  • Delivery runs on your own infrastructure with your own keys, logs, and task history.
  • The PR loop is GitHub-native and plain comments become follow-up commits.
  • You choose from Claude Code, Codex, Antigravity CLI, Mistral Vibe, and OpenCode per task.
  • Your existing provider subscriptions or API keys fund agent usage directly.

Tracker-hosted coding sessions versus a self-hosted GitHub PR loop.

QuestionLinear Coding SessionsProPR
Where the agent runsLinear's cloud sandboxYour server
Work starts fromA Linear issue, comment, or Slack threadA GitHub issue, plan, PR comment, CLI, or MCP
AgentsClaude Code or Codex, set per workspaceClaude Code, Codex, Antigravity CLI, Mistral Vibe, OpenCode, per task
BillingPrepaid AI credits: tokens plus sandbox timeYour provider plans or API keys; no markup
Opens a GitHub pull requestYesYes
Review comments become follow-up commitsYes: delegate review actions to the Linear AgentAny plain comment, any agent
Self-hosted infrastructure, keys, and logsNoYes

Verified November 2026: Linear coding sessions docs, Linear AI credits, coding environments changelog, Linear GitHub integration.