YouTrack MCP Server: Connect AI to Issues and Project Context
A YouTrack MCP server connects an AI assistant to your YouTrack issues, sprints, and project data through a standard tool call. Here is how to set up the connection, define what the AI can read and change, and carry shared rules and handoff notes across assistants while YouTrack stays as the issue tracker.
Updated 4 min read
A YouTrack MCP server connects an AI assistant to YouTrack issues, agile boards, and project data through a standard tool call. The assistant can read issue details, search by field or query, check sprint status, and update issue fields without leaving the coding or analysis session. Tools like fenbs carry the shared rules and handoff notes your team has agreed on across assistants, so YouTrack remains the source of truth for issues while fenbs holds the context the AI needs to work on them correctly.
YouTrack is available as YouTrack Cloud (opens in a new tab) and YouTrack Server. Both expose a REST API (opens in a new tab) that MCP servers use. Check the MCP server directory for available YouTrack MCP packages.
Connect an AI assistant to YouTrack
- In YouTrack, create a permanent token for the account the AI will use. Go to your profile, select Account Security, and generate a token with the required permissions. Read access is sufficient for most AI query tasks.
- Install and configure the YouTrack MCP server package. Set the YouTrack base URL and the permanent token as environment variables in the server config.
- Add the server to your AI client's MCP config. See MCP server configuration for the format.
- Restart the AI client and ask the assistant to search for open issues in a project you know well. Verify the issues it returns match what you see in YouTrack.
- Test update operations only if the MCP server exposes them and you have explicitly decided to allow write access. Start with a low-risk field, such as adding a comment or changing a tag.
What the AI can query in YouTrack
Through the YouTrack REST API, an AI assistant can search issues using YouTrack's query language, read issue fields (summary, description, state, assignee, priority, custom fields), list issues in a sprint or agile board, retrieve project metadata, and check saved searches. This covers the majority of status and planning questions a development team asks.
Write operations include creating issues, updating field values, adding comments, and transitioning states. These change the project record and should be used carefully. Require the AI to describe what it intends to change before making any update, so a team member can confirm.
Define access boundaries
The permanent token's permissions in YouTrack control what the AI can read and change. Create a dedicated token for AI access with the minimum permissions needed. If the current project only requires reading issues, use read-only permissions. Add write permissions separately if and when a specific use case requires them.
Document the permitted operations as a rule in the project context: which issue fields the AI may update, which state transitions it may trigger, and which operations always require a human to confirm. This makes the boundary visible to every session, not just the one where the decision was made.
Share rules and handoff notes across assistants
YouTrack tracks issues. It does not track the shared rules your team uses when working on them — the workflow conventions, the definition of done for each issue type, how bugs should be reproduced before fixing, which labels mean what. Those rules often live in a wiki, a document, or someone's head.
fenbs keeps those rules as Decisions and rules in the project, and the handoff notes — where the last session left off, what is in progress, what is waiting on a person — as Where we left off. When one coding assistant hands off to another, both read from the same fenbs project. YouTrack stays as the issue tracker; fenbs provides the context the AI needs to work on those issues well.
YouTrack has a built-in query language that is worth teaching to any AI assistant you connect. Rather than asking the assistant to browse issues manually, you can define common queries for it: project: {My Project} State: In Progress assignee: me for current work, or created: -1w Priority: Critical for recent critical issues. Record commonly used queries as context notes in fenbs so the assistant applies them consistently rather than constructing queries from scratch each session. The YouTrack search syntax documentation (opens in a new tab) covers the full query language, including field names, date ranges, and logical operators that work across both cloud and server deployments.
Carry shared rules and handoff notes across every coding session
Record your team's workflow conventions, definition of done, and issue-handling rules as Decisions and rules in fenbs. At the start of each session, the AI reads those rules before touching any YouTrack issues. Before the session ends, save where each task stands. Any connected assistant picks up from there.
Create a free accountFree for one person · No credit card · Works with Claude, ChatGPT, Cursor and more
How to do this with fenbs
- Create a free account and add a project that maps to the YouTrack project the AI is helping with.
- Record workflow conventions as Decisions and rules: how bugs are reproduced, which state transitions the AI may trigger, the definition of done for different issue types.
- Add the current in-progress tasks to the fenbs board. These are the tasks the AI is actively helping with, not a duplicate of YouTrack; they carry the planning context YouTrack does not.
- Connect your coding assistant to fenbs at Integrations. It reads the rules and current task status before querying YouTrack.
- Before switching assistants or ending a session, save where each task stands in fenbs so the next assistant starts from a known point.
fenbs does not replace YouTrack as the issue tracker. It carries the shared rules and handoff context that help the AI work on YouTrack issues correctly across sessions and tools.