# AI use cases for Thinkrail (JetBrains/thinkrail) Repo: [JetBrains/thinkrail](https://github.com/JetBrains/thinkrail) | Use cases: 3 | Page: https://dedupedai.com/use-cases/jetbrains-thinkrail/ ThinkRail is JetBrains' agentic IDE built around the pi coding agent. It cuts isolated git worktree workspaces so agents work in parallel without touching your checkout, wraps them in a real IDE (Monaco editor, file tree with git marks, terminals), keeps durable project knowledge in a graph of living SPEC.md files, reads your existing Agent Skills and connects MCP servers, with JetBrains AI or your own provider credentials. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Run parallel coding agents in isolated git worktrees: https://dedupedai.com/use-cases/jetbrains-thinkrail/run-parallel-agents-in-isolated-worktrees/ - Keep project intent alive across agent sessions with living specs: https://dedupedai.com/use-cases/jetbrains-thinkrail/keep-project-intent-in-living-specs/ - Review agent changes and ship the pull request from one window: https://dedupedai.com/use-cases/jetbrains-thinkrail/ship-a-pr-from-thinkrail/ ## How to run parallel coding agents in isolated git worktrees with Thinkrail Repo: [JetBrains/thinkrail](https://github.com/JetBrains/thinkrail) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Open a repo in ThinkRail and give each agent its own workspace and branch, then merge the best one. **Who it is for:** A developer who wants to try two or three approaches to a task at once without agents trampling each other. **The problem:** Several agents in one checkout overwrite each other's files, and juggling worktrees by hand is tedious. **The outcome:** Separate workspaces, each a git worktree with its own branch, where you compare results and merge the good branch. ### Steps 1. Install on macOS or Linux with `curl -fsSL https://raw.githubusercontent.com/JetBrains/thinkrail/main/install.sh | bash` 2. Open a repo with `thinkrail ~/code/my-repo` 3. Create two or more workspaces from the project; each becomes a git worktree with its own branch 4. Start a pi chat in each workspace with the approach you want it to try 5. Review the diffs, merge the good branch and delete the rest ### Prompt to paste into your agent ```text Open in JetBrains/thinkrail, create two workspaces, and have one agent implement with approach A and the other with approach B. Compare the diffs and test results, then tell me which branch to merge and why. ``` ### Caveat ThinkRail is early (0.x releases with stable and nightly channels), so expect changes between versions. Risk, remote script: The install command pipes install.sh from GitHub into bash; read it first if you prefer. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jetbrains-thinkrail/run-parallel-agents-in-isolated-worktrees/ ## How to keep project intent alive across agent sessions with living specs with Thinkrail Repo: [JetBrains/thinkrail](https://github.com/JetBrains/thinkrail) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Let the agent read and maintain a connected graph of SPEC.md files beside your code. **Who it is for:** A team lead whose agents forget design decisions every time a new session starts. **The problem:** Chat context disappears at the end of a session, so the next agent re-learns or contradicts earlier decisions. **The outcome:** A graph of SPEC.md files the agent searches and updates through spec tools, rendered as a tree in the workbench. ### Steps 1. Open your repo with `thinkrail ~/code/my-repo` 2. Use the bundled project setup workflow to start the spec graph 3. Ask the agent to record decisions and intent in SPEC.md files as it works 4. Browse the graph in the read-only Specs tool 5. Start a new chat later and check that the agent reads the specs before changing code ### Prompt to paste into your agent ```text In JetBrains/thinkrail, set up living specs for : write SPEC.md files for the main modules with their purpose and key decisions, link them into a graph, and keep them updated as you implement . ``` ### Caveat Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jetbrains-thinkrail/keep-project-intent-in-living-specs/ ## How to review agent changes and ship the pull request from one window with Thinkrail Repo: [JetBrains/thinkrail](https://github.com/JetBrains/thinkrail) | Difficulty: Beginner | Time: 20 minutes | Basis: From the README Leave anchored review comments, hand them to the agent and open the PR with its CI checks in view. **Who it is for:** A developer who wants to review agent work like a teammate's PR and merge it without switching tools. **The problem:** Reviewing agent diffs, pushing branches and watching CI usually means bouncing between terminal, editor and browser. **The outcome:** Agent changes reviewed with anchored comments, the branch pushed, a PR opened and CI and review status visible in the topbar. ### Steps 1. Work on a task in a ThinkRail workspace 2. Leave anchored review comments on files and diffs and send them to the chat as structured context 3. Let the agent address them, committing each completed plan step on its own 4. Use the topbar to publish, push and open a PR described from the plan and commits 5. Watch CI checks and the review decision, hand failing checks to the agent, then merge ### Prompt to paste into your agent ```text In JetBrains/thinkrail, review the changes in this workspace, leave anchored comments on anything unclear, have the agent fix them, then open a GitHub PR described from the plan and commits and report the CI status. ``` ### Caveat Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jetbrains-thinkrail/ship-a-pr-from-thinkrail/