Self-hosted Jenkins MCP alternatives
The Jenkins server already runs the way you want here: a local process that talks to a Jenkins controller you operate, listing jobs, triggering and stopping builds, and reading logs and test results. If you are changing tools but keeping the arrangement, the server and credentials on your own machine or network, the picks below install the same way.
Every server here runs locally over stdio against infrastructure you control. Three of them, the reference Fetch, Memory, and Time servers plus the Git server, are not CI platforms; they are local building blocks an agent pairs with a build tool. The notes are explicit about which is which.
The 8 best self-hosted alternatives
Running locally against a SonarQube instance you host, Sonar's official server analyzes code quality, security, and coverage. It inspects code rather than running builds, so it complements a self-hosted CI controller.
Set up SonarQube →Argo CD's server runs locally and manages GitOps applications: list and sync apps, read resource trees and workload logs, and run resource actions against a cluster you operate, deployment by reconciling Git state.
Set up Argo CD →Pulling build failure logs, finding flaky tests, validating config, and running pipelines, the official CircleCI server runs locally. The connector stays on your machine even though CircleCI itself is a managed platform.
Set up CircleCI →Buildkite's official server runs locally and reads pipelines, builds, jobs, logs, and artifacts and triggers builds, a fit for teams who keep CI agents on their own hardware and the orchestrator on Buildkite.
Set up Buildkite →Anthropic's reference Fetch server runs locally and converts a URL to clean markdown. It is not CI; pair it with a build server when a pipeline step needs to read an external page.
Set up Fetch →The reference Memory server runs locally and gives an agent a persistent knowledge graph of entities, relations, and observations on your disk, carrying state across CI sessions rather than running builds.
Set up Memory →The reference Time server runs locally and returns the current time and converts timezones from the IANA database, a small utility an agent uses around scheduled jobs.
Set up Time →Version control is the source side of a build, and the reference Git server runs locally on a repo path to read status, diffs, commits, branches, and history. It is not a build runner itself.
Set up Git →
How to choose
For a build-control connector that stays local, CircleCI and Buildkite keep the process on your machine while the platform runs elsewhere; Argo CD fits self-hosted GitOps delivery and SonarQube analyzes your code. Git, Fetch, Memory, and Time are not CI engines at all, they are local primitives an agent uses for source, web reads, state, and timing alongside whichever build tool you run.
FAQ
- Is the Jenkins MCP server self-hosted?
- Yes. It runs as a local process and connects to a Jenkins controller you operate, so the server and its credentials stay on your own infrastructure. The alternatives here also ship a local stdio command, though some connect outward to a managed CI platform.
- Why are Git, Fetch, Memory, and Time included as Jenkins alternatives?
- They are not replacements for Jenkins. They are reference and utility servers that run locally and give an agent version control, web fetching, a persistent knowledge graph, and time lookups, the primitives a CI agent commonly uses for individual steps rather than for running pipelines.