Docs · AI assistants

Connect Claude to Sprint Track, and stay in control.

Sprint Track runs a remote MCP server, so Claude on the web, desktop, and mobile, Claude Code, and other MCP clients can work in your workspace. Reads happen directly. Edits wait for an approver in Sprint Track unless your workspace auto-approves them, and sensitive and destructive changes always do.

OAuth sign-in, no API key Never more than your role Revoke any time

Jump to

How it works Requirements Claude apps Claude Code Other MCP clients What you can ask Approvals and safety Privacy

How it works

One connection, one workspace, your permissions.

Sprint Track speaks the Model Context Protocol (MCP), the open standard AI assistants use to connect to other apps.

Server URL

https://api.sprintrack.com/mcp. It is the only thing you need to paste. There is no API key, client ID, or secret to copy.

Sign-in

OAuth. You sign in to Sprint Track in your browser and approve the connection. Claude registers itself automatically.

Workspace

Each connection is bound to the one workspace you choose while connecting. To work in a different workspace, revoke the connection under AI assistants, then Connections, and connect again.

Access

Choose Full access, which is everything your Sprint Track role allows, or Custom, to pick exactly which capabilities, and optionally which projects, the assistant gets. A connection never goes beyond your role.

Changes

Reads run straight away. Edits wait for an approver in Sprint Track, who is notified, unless your workspace auto-approves routine edits from this app. Sensitive and destructive changes always wait for approval.

Audit

Everything an assistant does goes through the same permission checks and audit trail as the Sprint Track app.

Requirements

Before you connect.

What your workspace, your role, and your Claude plan need.

Package

Your workspace needs a package that includes AI assistant access: Launch, AI, or Enterprise. It is not included on Free.

Role

Owners, Admins, Members, and Viewers can connect by default. Guests cannot. A custom role needs the permission to connect the workspace to ChatGPT, Claude, and other MCP clients.

Claude plan

Custom connectors are available on Claude's Free, Pro, Max, Team, and Enterprise plans. Free is limited to one custom connector. On Team and Enterprise, an organization Owner adds the connector first.

Workspace settings

Your administrators can turn each AI app on or off, limit what it may do, choose whether its routine edits are auto-approved, and block unverified clients. All are described under Approvals and safety.

Claude on the web, desktop, and mobile

Add Sprint Track as a custom connector.

Add it once. The connector then works on claude.ai, in Claude Desktop, and in the Claude mobile apps, because the connection runs from Anthropic's servers.

Open Customize, then Connectors
On claude.ai or in Claude Desktop. Some versions label this Settings, then Connectors.
Click +, then Add custom connector
3
Paste the Sprint Track server URL
Name it Sprint Track. Leave Advanced settings empty; no client ID or secret is needed.
Server URL
```
https://api.sprintrack.com/mcp
```
Click Connect and sign in to Sprint Track
Claude opens Sprint Track in your browser. Sign in the way you normally do.
Choose a workspace and the access to grant
Pick one workspace, then Full access or Custom. Approve to return to Claude.
Start asking
In a chat, check that Sprint Track is switched on in the tools menu, then ask about your work.

Claude Team and Enterprise

An organization Owner first opens Organization settings, then Connectors, clicks Add, then Custom, then Web, and enters the same server URL. Each member then opens Customize, then Connectors, and clicks Connect to sign in with their own Sprint Track account.

Claude Code

Add the server from your terminal.

Claude Code signs in with the same OAuth flow. Nothing is stored in your project files.

1
Add the server
Add --scope user to make it available in every project, not only the current one.
Terminal
```
claude mcp add --transport http sprint-track https://api.sprintrack.com/mcp
```
2
Authenticate
Start Claude Code, run /mcp, select sprint-track, and choose Authenticate. Your browser opens so you can sign in, choose a workspace, and grant access.
In Claude Code
```
/mcp
```
Check the connection
Run /mcp again and sprint-track shows as connected. The Sprint Track workflow prompts, such as plan_sprint and standup, appear as slash commands.

Or install the Sprint Track plugin

The plugin adds the same server plus a Sprint Track skill that teaches Claude the approval rules and common workflows, such as sprint planning and triage.

1
Add the Sprint Track marketplace
In Claude Code
```
/plugin marketplace add axceera/sprint-track
```

Install the plugin

In Claude Code

/plugin install sprint-track@sprint-track

Authenticate
Run /mcp, select the Sprint Track server, and choose Authenticate.

Cursor, VS Code, and other MCP clients

Any MCP client that supports OAuth can connect.

Add a remote MCP server that uses HTTP (Streamable HTTP) with the same URL. The first time the client connects, it opens your browser so you can sign in, choose a workspace, and grant access.

Cursor: ~/.cursor/mcp.json

mcp.json

{
  "mcpServers": {
    "sprint-track": {
      "url": "https://api.sprintrack.com/mcp"
    }
  }
}

VS Code: .vscode/mcp.json

mcp.json

{
  "servers": {
    "sprint-track": {
      "type": "http",
      "url": "https://api.sprintrack.com/mcp"
    }
  }
}

Other clients: add a remote server of type HTTP with the URL https://api.sprintrack.com/mcp, then complete the sign-in when prompted. No API key is needed.

What you can ask

Plan, triage, report, and update work in plain language.

The assistant can do anything your role allows within the access you granted. A few examples:

Find and summarize

"What is overdue in the Mobile project, and who owns it?"

Triage

"Find backlog issues with no priority or owner and suggest both. Show me a table before changing anything."

Plan a sprint

"Plan a two-week sprint for the Website project from the top of the backlog."

Update work

"Move PM-42 to In Progress, assign it to Maya, and add a comment with this update."

Report

"Write a status report for leadership on the current cycle."

Standup

"Prepare my standup from what I worked on yesterday."

Knowledge

"Find the launch brief and summarize the open decisions."

Time

"Log two hours on PM-42 for yesterday."

Approvals and safety

Edits wait for an approver. You decide what runs on its own.

The assistant works through the same permission checks as the Sprint Track app, and it cannot approve anything on your behalf.

Reads

Run directly, within your role and the access you granted.

Routine edits

Creating and updating issues, comments, cycles, labels, time logs, and similar records wait for an approver by default. Administrators can auto-approve them for each app on the AI assistants page; they are recorded in the audit trail either way.

Sensitive and destructive changes

Deleting or revoking anything, changing membership, rates, sharing, or project scope, running agents and imports, and every GitHub write create an approval request instead. The assistant gives you a link, you review the exact change in Sprint Track, and it runs as soon as you approve it. An assistant can never approve its own request.

Per-app settings

Workspace administrators can turn Claude, ChatGPT, or other clients on or off, limit each to read only or read and write, and choose whether its routine edits are auto-approved.

Unverified clients

Administrators can block MCP clients that Sprint Track cannot verify. Claude on the web, desktop, and mobile is verified. Other clients, which can include editors and command-line tools, cannot connect while this is on.

Revoke any time

Open AI assistants, then Connections, in Sprint Track to see each connection, narrow its access, or revoke it. Administrators can review and revoke every member's connections.

Untrusted content

Sprint Track instructs connected assistants to treat issue text, comments, and documents as data, never as instructions. Review what an assistant proposes before you approve it.

Privacy

Where your data goes.

Connecting an AI assistant shares data with that assistant's provider, and only because you asked it to.

Your choice

Nothing is shared with an AI provider until you connect one. Connecting is opt-in and started by you.

What is shared

The workspace, project, and issue data the assistant reads or changes for your requests is sent to the AI provider you connected, such as Anthropic for Claude.

Provider terms

Once that data reaches the provider, it is handled under your agreement with them and their privacy policy.

Stopping

Revoke the connection in Sprint Track, or remove the connector from your assistant. Revoking ends the assistant's access straight away.

Anthropic (for Claude) and OpenAI (for ChatGPT) are listed on our sub-processors page. See our Privacy Policy and security overview for more.

Need a hand connecting?

Email hello@sprintrack.com with the assistant you are using and what you see, and we will help you get connected. Using ChatGPT? See Connect ChatGPT to Sprint Track.

Connect your workspace when we open.

ChatGPT and Claude connect on Launch and above. Join the waitlist and we will email you when Sprint Track opens.

We will email you when Sprint Track opens.

See pricing