> ## Documentation Index
> Fetch the complete documentation index at: https://developers.everhour.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Everhour MCP server

> Connect AI assistants like Claude and ChatGPT to your Everhour account to track time, manage tasks, and read timesheets in plain language.

The Everhour MCP server lets an AI assistant read and write your Everhour data on your behalf — start a timer, log time, check your timesheet, or find a task — without you switching tabs or memorizing IDs. You describe what you want in plain language and the assistant calls the right tool.

It implements the [Model Context Protocol](https://modelcontextprotocol.io) (MCP), an open standard for connecting AI clients to external tools and data. Any MCP-capable client can connect.

## Server details

<CardGroup cols={2}>
  <Card title="Endpoint" icon="link">
    `POST https://api.everhour.com/mcp`
  </Card>

  <Card title="Transport" icon="arrows-left-right">
    Streamable HTTP
  </Card>

  <Card title="Protocol" icon="code-branch">
    MCP `2025-06-18`, JSON-RPC 2.0
  </Card>

  <Card title="Authentication" icon="key">
    OAuth 2.1 (or `X-Api-Key` for manual setups)
  </Card>
</CardGroup>

## What you can do

Once connected, your assistant has access to 24 tools covering everyday time tracking:

* **Track time in real time** — start, stop, and check your running timer.
* **Log and edit time** — add entries (in seconds), fix a duration, delete a wrong entry, or log time retroactively.
* **Review timesheets** — read your weekly timesheet with daily totals and a per-task breakdown.
* **Find work across tools** — search tasks, list recently tracked tasks, and resolve a task ID or URL from Asana, Jira, GitHub, Linear, and other integrations to the matching Everhour task.
* **Read projects, clients, and team** — list and fetch projects, clients, and team members.
* **Report on the team** — pull hours by project or member, project budgets, time off, and weekly capacity over any date range (subject to your role).

See the [tools reference](/mcp-tools) for the full list with parameters.

## Authentication

Clients connect with [**OAuth 2.1**](/oauth). You add the server URL and approve access in your browser — no app to create and no credentials to copy, because Everhour registers your client automatically. The access token is short-lived and refreshes on its own. See [Connect an MCP client](/mcp-connect).

For clients that can't do OAuth, or for headless automation, you can authenticate with an Everhour API key in an `X-Api-Key` header instead. The key is long-lived and grants full account access, so prefer OAuth where it's available.

<Note>
  Whichever method you use, the assistant can only do what your Everhour role allows — the same permissions that govern the REST API.
</Note>

## Requirements

* An active Everhour account.
* An MCP-capable AI client (see [supported clients](/mcp-connect)).
* To authenticate: either a client that supports OAuth (no setup — it registers itself), or your [Everhour API key](https://app.everhour.com/#/account/profile) for the header-based method.

## Next steps

<CardGroup cols={2}>
  <Card title="Connect an MCP client" icon="plug" href="/mcp-connect">
    Set up Claude, ChatGPT, Cursor, and other clients.
  </Card>

  <Card title="Tools reference" icon="wrench" href="/mcp-tools">
    Every tool the assistant can call, with parameters.
  </Card>
</CardGroup>