YaTracker MCP Server

How to use

You connect to the MCP server from your MCP client using the configured endpoint. For remote interactions, you typically reference the HTTP endpoint, while for local, self-contained setups you run the server locally and point your MCP client to the local URL. The available operations let you search and list issues, read and add comments, manage transitions, handle attachments, and access queue metadata. Use the client to list issues in a queue, fetch details for a specific issue, update fields, add comments or attachments, and download attachments. You can also list queues, fields, and workflows to understand how work items are modeled in your Yandex Tracker instance.

To start quickly, choose Docker for a straightforward setup, then configure credentials in the environment file and bring the stack up. If you prefer a local Python setup, install dependencies, create the environment file, and run the CLI-based server locally. The MCP server exposes an endpoint suitable for standard MCP clients, and you can configure the client to connect via the provided URL.

How to install

Prerequisites you need before installing: Docker (optional but recommended), Python 3.11 or newer, and access to a Yandex Tracker API token with your Organization ID.

Option 1 — Docker (recommended) 1. Create an environment file from the sample 2. Edit the file to include your credentials 3. Start the services with Docker Compose 4. The MCP server will be available on port 8080.

# 1. Create .env from the sample
cp config/sample.env .env
# Edit .env to add credentials like YATRACKER_TOKEN and YATRACKER_ORG_ID

# 2. Run the stack
docker-compose up -d

Additional configuration and notes

Prerequisites and credentials you will configure in the environment file include the Yandex Tracker API token and your organization ID. The sample environment file shows the required variables you must set.

Option 2 — Local installation 1. Install dependencies 2. Create the environment file 3. Use the CLI to start and interact with the server

# 1. Install dependencies
pip install -r requirements.txt

# 2. Create environment file
cp config/sample.env .env
# Add YATRACKER_TOKEN and YATRACKER_ORG_ID

# 3. Run via CLI
PYTHONPATH=src python3.11 -m scripts.cli issues list --queue CRM

MCP configuration example for Claude Code

If you are wiring a Claude Code MCP client, configure the MCP server endpoint as shown.

{
  "mcpServers": {
    "yatracker": {
      "transport": "sse",
      "url": "http://localhost:8080/sse"
    }
  }
}

yatracker_search_issues

Search and list issues across queues with filters and limits.

yatracker_get_issue

Retrieve detailed information for a specific issue by key.

yatracker_create_issue

Create a new issue or lead in a specified queue.

yatracker_update_issue

Update fields on an existing issue.

yatracker_list_comments

List comments for a given issue.

yatracker_add_comment

Add a new comment to an issue.

yatracker_add_comment_with_attachment

Add a comment together with an attachment to an issue.

yatracker_list_transitions

List available status transitions for an issue.

yatracker_execute_transition

Execute a status transition on an issue.

yatracker_download_attachments

Download attachments from an issue.

yatracker_attach_file

Attach a file to an issue.

yatracker_list_queues

List all available queues.

yatracker_list_queue_fields

List fields for a specific queue.

yatracker_list_all_fields

List all system fields.

yatracker_list_queue_issue_types

List issue types for a queue.

yatracker_get_queue_workflows

Get the workflow graph for a queue.