Trunk Platform

Eliminate merge bottlenecks and flaky test failures at scale

Modern CI is fast. But scale breaks everything else.

Your build system (Bazel, Nx, Gradle) gives you massive parallelization and smart caching. But at high velocity, two critical reliability problems emerge:

Merge queue bottlenecks

Without a merge queue, PRs break main when independently-passing changes conflict.

With a traditional merge queue, you've solved stability—but created a serial processing bottleneck. PRs test one-at-a-time in a single queue, so frontend changes wait behind unrelated backend tests. Queue wait times grow linearly with PR volume.

Flaky test multiplication

At tens of thousands of tests, even 1% flake rates mean false failures on every test run. In merge queues, each flaky failure blocks multiple PRs—turning minor issues into major bottlenecks.

Worse, flaky tests erode trust: When developers know there's a chance a re-run will pass, they stop trusting test failures altogether.

Trunk eliminates both bottlenecks: Graph-based parallel merge lanes that test non-overlapping changes simultaneously (while guaranteeing main stability), plus automated flaky test quarantine that prevents false failures from blocking your pipeline.

Trusted by

Features

Advanced Merge Queue

Enterprise-grade merge queue that accelerates PR throughput while protecting your main branch. Purpose-built for high-velocity monorepos with intelligent batching to reduce CI costs by up to 90%, parallel queues for testing independent code changes simultaneously, and anti-flake protection to prevent transient failures from blocking merges.

Learn more about Merge Queue

Automated Flaky Test Management

Trunk Flaky Tests automatically detects flaky tests and quarantines them to prevent CI failures. Track, manage, and fix flaky tests across any language, environment, or test framework.

Learn more about Flaky Tests

Agent-Powered Root Cause Analysis (beta)

Automatically identifies and explains the root cause of test and CI failures using autonomous AI agents. Delivers actionable insights directly in GitHub PR comments and Slack, including failure classification, impact assessment, and suggested fixes. Also available in Cursor and VSCode via the Trunk MCP Server.

CI Autopilot is currently in private beta. Sign up for the waitlist to be the next to get access.

Learn more about CI Autopilot

Get Started

Last updated