MCP Server Circleci logo

MCP Server Circleci

by CircleCI

Connects AI clients to CircleCI for pipeline diagnostics, test analysis, configuration checks, reruns, deployments, and usage review.

Install MCP Server Circleci

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "circleci-mcp-server-circleci": {
      "command": "npx",
      "args": [
        "-y",
        "@circleci/mcp-server-circleci@latest"
      ],
      "env": {
        "CIRCLECI_TOKEN": "YOUR_VALUE_HERE",
        "CIRCLECI_BASE_URL": "YOUR_VALUE_HERE",
        "MAX_MCP_OUTPUT_LENGTH": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

Install not yet sandbox-tested— the server itself is confirmed live (we called its tools/list endpoint). The stdio install command is a separate, ongoing check.

Overview

CircleCI/mcp-server-circleci MCP server exposes CircleCI project, pipeline, job, test, artifact, configuration, usage, and rollback operations through MCP. It runs locally over stdio with NPX or Docker, using a CircleCI personal API token, and supports project slugs, CircleCI URLs, or workspace metadata for project selection. Reach for it when an agent needs to inspect failures, validate configuration, trigger or rerun workflows, or review CircleCI usage. The package is deprecated and should be replaced with CircleCI’s hosted MCP server or CircleCI CLI MCP.

Use cases

  • Inspect failed CircleCI build logs
  • Check pipeline and test status
  • Validate and repair CircleCI configuration
  • Trigger or rerun CircleCI workflows
  • Review usage and underused resource classes

Key features

  • Retrieve build failure logs and job test results
  • Analyze flaky tests and CircleCI configuration
  • Check, trigger, and rerun pipelines
  • List projects, artifacts, and component versions
  • Download usage data and inspect resource utilization
  • Run project rollback workflows

Tools13

What MCP Server Circleci lets your agent do. Expand a tool for its parameters.

Verified live~12.5k tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked Today

Recent health check succeeded.

28/28 checks healthy over the last 93d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 9d ago via OSV.dev · @circleci/mcp-server-circleci@latest (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars94Stargazers on the source repository.
npm downloads80kPackage downloads in the last 30 days.
Last commit20d agoMost recent push to the default branch.
Tools exposed13Callable tools this server registers over MCP.
Directory activity2 viewsConfig copies, upvotes, and views on AllMCPs.

How MCP Server Circleci works

What CircleCI/mcp-server-circleci does

CircleCI/mcp-server-circleci connects MCP-compatible clients to CircleCI. Its tools cover CI troubleshooting, pipeline control, test and artifact inspection, configuration review, deployment version lookup, rollback workflows, and usage analysis.

For diagnostics, agents can retrieve build failure logs, inspect job test results, find flaky tests, check the latest pipeline state, and list artifacts. Configuration analysis accepts the complete CircleCI YAML content and returns validation errors or guidance. Pipeline operations can start a pipeline, rerun a workflow from the beginning or from its failed job, and trigger rollback procedures when the project has the required rollback setup.

The server also lists followed projects and component versions. Usage-related tools download organization usage data as a CSV and identify jobs whose CPU or RAM resource classes are below a selected utilization threshold.

How it works

The local server communicates with an MCP client through stdio. CircleCI operations authenticate with a personal API token. Most project-oriented tools accept exactly one of three identification approaches: a project slug with a branch, a CircleCI URL supplied by the user, or workspace root, Git remote URL, and branch information. Workflow reruns instead accept a workflow ID or workflow URL.

A common sequence starts with list_followed_projects, then uses the returned project slug for a status check, log lookup, test query, or pipeline action. Tools that select component versions can require several calls: first choose an environment, then a component, and finally a version. Pipeline triggering may also require a second call when a project exposes multiple pipeline definitions.

Large log and test responses can be truncated. The tool output includes an MCP truncation marker when this happens, and agents are instructed to acknowledge the warning before analyzing the remaining data. Full build logs can be written to an explicit output directory.

Setup and configuration

The package can be launched with NPX or Docker. NPX installation requires a supported Node.js release in the documented ranges and pnpm. Docker is an alternative for local deployments. A CircleCI personal API token is required and is passed through CIRCLECI_TOKEN.

Example NPX configuration:

config.json
{
  "mcpServers": {
    "circleci-mcp-server": {
      "command": "npx",
      "args": ["-y", "@circleci/mcp-server-circleci@latest"],
      "env": {
        "CIRCLECI_TOKEN": "your-circleci-token",
        "CIRCLECI_BASE_URL": "https://circleci.com",
        "MAX_MCP_OUTPUT_LENGTH": "50000"
      }
    }
  }
}

CIRCLECI_BASE_URL is optional for CircleCI cloud and is intended for on-premises customers. MAX_MCP_OUTPUT_LENGTH sets the maximum MCP response length and defaults to 50,000. Cursor, VS Code, and Claude Desktop configuration examples are included for both NPX and Docker.

Limitations and notes

This package is deprecated, no longer receives feature work, and is intended to be replaced by CircleCI’s hosted MCP server or CircleCI CLI MCP. The repository is scheduled for archival, although existing npm versions remain installable. Operating the self-managed remote transport requires special care because it brokers an organization’s CircleCI token; CircleCI’s hosted service is the direct replacement for that mode.

Tool calls require complete parameter sets and should not guess project slugs or URLs. URLs used as inputs must come from the user. Rollback actions depend on project-specific rollback pipeline configuration, deployment markers, component versions, and explicit confirmation. Usage CSV analysis also requires an absolute file path and the expected utilization columns.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about MCP Server Circleci

CircleCI/mcp-server-circleci is an MCP server that connects AI clients to CircleCI through a personal API token. Its main tools inspect pipeline status, failure logs, tests, artifacts, and configuration, while also triggering pipelines, rerunning workflows, analyzing usage, and supporting rollback operations. The package is deprecated.

Install it by configuring an MCP client to run `npx -y @circleci/mcp-server-circleci@latest`, or launch the Docker image instead. Set `CIRCLECI_TOKEN` in the client environment. Cursor, VS Code, and Claude Desktop configuration examples are available for both local methods.

Yes, it works with Cursor through a local stdio configuration using NPX or Docker. Add the server command and set `CIRCLECI_TOKEN`; `CIRCLECI_BASE_URL` and `MAX_MCP_OUTPUT_LENGTH` are optional settings. The package is deprecated, so new deployments should use CircleCI’s hosted or CLI MCP.

Use `get_latest_pipeline_status` with either a project slug and branch, a user-supplied CircleCI URL, or workspace root, Git remote URL, and branch. A typical workflow first calls `list_followed_projects` and then passes the exact returned project slug to the status tool.

Use `rerun_workflow` with the workflow ID or a workflow URL, and set `fromFailed` to true to continue from the failed job. Set it to false to rerun from the start. The workflow identifier or URL must be supplied rather than guessed.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewMCP Server Circleci AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/circleci-mcp-server-circleci?style=directory)](https://allmcps.com/mcp/circleci-mcp-server-circleci)
HTML Embed
<a href="https://allmcps.com/mcp/circleci-mcp-server-circleci"><img src="https://allmcps.com/api/badge/circleci-mcp-server-circleci?style=directory" alt="MCP Server Circleci on AllMCPs" /></a>