Mergetrain MCP Server

mergetrain

Parallel agents need a serial integration spine.

mergetrain is a local-first deploy train for coding-agent worktrees. Agents commit and enqueue their branches; one runner assembles them in order, tests the combined tree, and atomically updates your Git refs only after explicit approval.

The problem

Worktrees let several agents edit one repository without sharing a checkout. They do not decide landing order, test the combined result, prevent push races, or tell you what happened if a laptop dies mid-push.

Without an integration boundary, the human becomes that boundary: rebase every finished branch, rerun gates after each merge, resolve cross-branch failures, and decide which session may push. The parallel coding gain disappears at the last mile.

mergetrain makes that last mile a durable protocol:

agent branches → FIFO queue → isolated integration worktree → combined gates
               → explicit approval → one atomic push → post-push verification

Who should use it?

Use mergetrain when:

• multiple coding agents finish branches in the same repository throughout the day;
• agents work in Git worktrees and should enqueue rather than push deploy refs;
• the combined result must pass local tests before it lands;
• you want unattended processing only for explicitly pre-approved jobs; or
• one local hub should show queues and runners across several repositories.

It is harness-agnostic: Codex, Claude Code, scripts, and humans all use the same CLI and JSON contract.

Who should not use it?

You probably do not need mergetrain when:

• one person or agent lands one branch at a time;
• every change already goes through a PR and your forge-native merge queue;
• you need a hosted review UI, organization-wide permission system, or remote runner service; or
• you are looking for a general job queue, CI provider, or deployment platform.

For PR-first teams, use GitHub Merge Queue or GitLab Merge Trains. mergetrain is for local-agent, worktree-first integration, with or before a PR.

See it in 60 seconds

uvx mergetrain demo

The demo creates a disposable repository and local bare remote, then runs four real branches through FIFO merge, a combined-only gate failure, conflict attribution, and deployment of the compatible train. Use --keep to inspect the result afterward.

Install and first run

# Install the machine-level CLI
uv tool install mergetrain          # or: pipx install mergetrain
# macOS: brew install yongjip/tap/mergetrain

cd /path/to/your/repo

# Write .mergetrain.yaml plus agent instructions
mergetrain init --project my-app --write

# After an agent commits its task branch
mergetrain enqueue \
  --task "add health check" \
  --branch agent/health \
  --capture-sha

# Inspect first, then validate or deploy explicitly
mergetrain status --json
mergetrain run-batch --validate-only
mergetrain run-batch --deploy

deploy means the configured atomic Git ref update; it does not imply an App Store, Kubernetes, or other provider release. Projects can select integrate or push terminology without changing the stable machine contract.

mergetrain init also writes agent-facing instructions. The essential rule is simple: agents commit and enqueue; one runner owns merge → test → push → verify. Unattended daemons process only jobs that a human explicitly enqueued with --auto.

See the quickstart for configuration, dashboard, daemon, and multi-repository Hub setup.

Why not just worktrees and git merge?

Worktrees solve parallel editing. mergetrain solves serialized integration.

Plain worktrees remain the execution lanes. mergetrain is the spine that joins their results without turning the operator into a merge coordinator.

Why not GitHub or GitLab merge queues?

They solve a related problem for a different operating model.

The models can coexist: push a validated train to a review branch and open one PR, or reserve individual PRs for changes that need discussion. The PR workflow guide covers direct, one-PR, split-PR, and validation-only patterns.

Core safety guarantees

• Exact train identity. Approval names the task HEADs and integration base; changed branches or a moved base cannot silently reuse that approval.
• Combined gates before push. A green branch is not enough. The assembled train passes the configured gates, or nothing lands.
• One fenced owner. SQLite claims and lease tokens prevent concurrent or stale runners from mutating the same train.
• Atomic remote update. Payload refs and a permanent refs/mergetrain/deploys/<sha> recovery ref update together.
• Remote-truth recovery. Write-ahead markers and pinned commits let reconcile/recover determine whether a killed push landed, without replaying a successful deploy or calling a missing one shipped.
• Explicit automation. A bare run never deploys. Daemons touch only pre-approved --auto jobs, and MCP deploy still requires attributable human confirmation.
• Observable state. doctor, status, inspection, events, and statistics expose structured state and the next safe action instead of asking an agent to infer it from processes or prose.

Queue state, locking, train assembly, and gates stay local. Your configured Git remote and post-push verification may still use external services. Gate and verify commands are trusted code; review the security boundary before enabling unattended jobs.

These guarantees are exercised on macOS and Linux across Python 3.10–3.14 and on Windows, including real-Git fault injection around git push --atomic. A dedicated soak repository completed 20 landed trains at a 100% land rate, including planned conflict recovery and a real killed-push reconciliation whose verdict matched the remote. See the soak evidence, then use mergetrain stats --json to inspect evidence from your own queue.

Go deeper

• Start: Quickstart · Install · CLI · Config
• Understand: Design and architecture · Machine contract · PR workflow comparison
• Operate: Efficient operation · Failure modes and recovery · Daemon · Multi-repo Hub
• Trust and extend: Security · Agent contract · Agent adoption benchmark · MCP server · Adapter pattern · Product scope

Status

The current release is v1.4.1. Machine-contract major 1 is additive-only: existing JSON keys are not removed or renamed without a contract-version change. Issues and operating reports are welcome on GitHub.

License

Released under the MIT License.