Yandex Tracker

You run a dedicated MCP server that bridges AI assistants with Yandex.Tracker, exposing a standardized set of actions to manage tasks, comments, projects, and users via a consistent MCP interface. This server runs locally or in your environment and handles authentication, rate limits, and data validation while translating MCP calls into Yandex.Tracker API requests.

How to use

To use this MCP server, you run it as a local service that your MCP clients can connect to via stdio. The server provides a set of tools to manage tasks, comments, queues, and users in Yandex.Tracker. Your MCP client can create, read, update, and search tasks, add and fetch comments, work with queues and user profiles, and run prompts to analyze or summarize tasks.

How to install

Prerequisites: Node.js 18+ and npm installed on your system.

1. Install the MCP server package globally.

2. Create a configuration file for the MCP server connection and environment.

3. Run the MCP server and expose it for MCP clients to connect.

Configuration and usage notes

Authentication uses a OAuth token passed via the Authorization header. You also provide your Yandex.Tracker cloud organization ID in the request header. The base API URL defaults to https://api.tracker.yandex.net/v2, but you can override it if needed.

Environment variables you will configure include the token, organization ID, base URL, log level, request timeout, and rate limit settings. These controls are used by the MCP server to securely connect to Yandex.Tracker and manage request pacing.

Below is a representative example of the MCP configuration block you would place in your environment to start the server.

{
  "mcpServers": {
    "yandex_tracker": {
      "command": "yandex-tracker-mcp-server",
      "env": {
        "YANDEX_TRACKER_TOKEN": "your_token_here",
        "YANDEX_TRACKER_CLOUD_ORG_ID": "your_org_id",
        "YANDEX_TRACKER_BASE_URL": "https://api.tracker.yandex.net/v2",
        "LOG_LEVEL": "info",
        "REQUEST_TIMEOUT": "30000",
        "RATE_LIMIT_REQUESTS": "100",
        "RATE_LIMIT_WINDOW": "60000"
      }
    }
  }
}

Tools and capabilities overview

You can perform the following actions through MCP calls. Each tool corresponds to a concrete operation against Yandex.Tracker.

Available tools and endpoints

• create_issue: Create a new task with summary, description, queue, type, priority, assignee, components, and labels. Returns the created task with its key.

• get_issue: Retrieve full information about a task by its key.

• update_issue: Update an existing task with provided fields and return the updated task.

• search_issues: Find tasks using a Tracker query language, with sorting, limit, and offset.

• transition_issue: Change a task status with an optional comment.

• add_comment: Add a comment to a task, with optional assignees.

• get_comments: Retrieve comments for a task with pagination.

• get_queues: List queues with optional expansion and filtering.

• get_queue: Get detailed information about a queue.

• get_myself: Retrieve information about the currently authenticated user.

• search_users: Find users by query with optional limit.

Practical tips for operators

• Protect your OAuth token and only expose it through environment variables in their runtime environment.

• Respect Yandex.Tracker rate limits by enabling the built-in backoff and queuing strategies.

• Validate all inputs to MCP calls to prevent malformed requests from reaching the API.

• Use the provided logging to monitor errors and performance, and avoid leaking tokens in logs.

Security and best practices

• Tokens should never be embedded in code or leaked into logs. Use environment variables to supply tokens.

• Implement proper error handling so clients can gracefully recover from transient failures.

Troubleshooting

• If a token expires, the server should log a clear message and prompt for a refreshed token.

• If you encounter 401 or 403 responses, verify your token and organization ID headers.

Available tools

create_issue

Create a new tracker task with summary, description, queue, type, priority, assignee, components and labels. Returns the created task with its key.

get_issue

Fetch full information about a task by its key.

update_issue

Update an existing task with provided fields and return the updated task.

search_issues

Search for tasks using a Tracker query with sort, limit and offset.

transition_issue

Change the status of a task and optionally add a comment.

add_comment

Add a comment to a task with optional assignees.

get_comments

Retrieve comments for a given task with pagination.

get_queues

List available queues with optional expansion and filtering.

get_queue

Get detailed information about a specific queue.

get_myself

Return information about the currently authenticated user.

search_users

Find users by a query with an optional limit.