gitlab

Overview

The gitlab MCP server is a Model Context Protocol (MCP) server that enables AI assistants and agents to interact directly with GitLab projects and resources through a structured, AI-friendly interface. It allows AI-driven workflows to work with repositories, issues, merge requests, and project metadata without switching tools or manually navigating the GitLab UI.

This server is ideal for development, code review, CI/CD visibility, and project management workflows that rely on GitLab as the system of record.

Transport

streamable-http

Tools

• merge_merge_request
• create_or_update_file
• search_repositories
• create_repository
• get_file_contents
• push_files
• create_issue
• create_merge_request
• fork_repository
• create_branch
• get_merge_request
• get_merge_request_diffs
• list_merge_request_diffs
• get_branch_diffs
• update_merge_request
• create_note
• create_merge_request_thread
• mr_discussions
• update_merge_request_note
• create_merge_request_note
• get_draft_note
• list_draft_notes
• create_draft_note
• update_draft_note
• delete_draft_note
• publish_draft_note
• bulk_publish_draft_notes
• update_issue_note
• create_issue_note
• list_issues
• my_issues
• get_issue
• update_issue
• delete_issue
• list_issue_links
• list_issue_discussions
• get_issue_link
• create_issue_link
• delete_issue_link
• list_namespaces
• get_namespace
• verify_namespace
• get_project
• list_projects
• list_project_members
• list_labels
• get_label
• create_label
• update_label
• delete_label
• list_group_projects
• list_wiki_pages
• get_wiki_page
• create_wiki_page
• update_wiki_page
• delete_wiki_page
• get_repository_tree
• list_pipelines
• get_pipeline
• list_pipeline_jobs
• list_pipeline_trigger_jobs
• get_pipeline_job
• get_pipeline_job_output
• create_pipeline
• retry_pipeline
• cancel_pipeline
• list_merge_requests
• list_milestones
• get_milestone
• create_milestone
• edit_milestone
• delete_milestone
• get_milestone_issue
• get_milestone_merge_requests
• promote_milestone
• get_milestone_burndown_events
• get_users
• list_commits
• get_commit
• get_commit_diff
• list_group_iterations
• upload_markdown
• download_attachment

Key Capabilities

• Code exploration and analysis — Browse and search source code across GitLab projects.
• Issue and merge request management — Create, update, and reason about issues and MRs programmatically.
• Project insight — Access repository metadata, commit history, and project structure.
• Cross-project discovery — Search across groups or projects to find related code or context.
• Workflow automation — Integrate GitLab-based actions into larger AI-driven development flows.

How It Works

The gitlab MCP server runs as an MCP service and connects to GitLab using configured credentials with appropriate scopes. AI clients communicate with the server over the MCP protocol, requesting repository context or project actions as part of broader reasoning workflows.

The server mediates all interaction with GitLab, handling authentication, API communication, and response normalization. Results are returned in structured formats that AI assistants can reason over directly, while ensuring that all access respects GitLab’s permission model.

By abstracting GitLab’s APIs behind MCP, this server makes GitLab resources feel like built-in tools for AI assistants — enabling workflows such as repository exploration, merge request review, and project planning to happen naturally inside AI-driven environments.