Astravue MCP server
{
"mcpServers": {
"astravue": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}{
"mcpServers": {
"astravue": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}{
"mcpServers": {
"astravue": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}{
"servers": {
"astravue": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}{
"mcpServers": {
"astravue": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.0.1
What it does
The Astravue MCP Server connects Astravue project management workspaces to AI tools like Claude, allowing you to search projects and tasks, create and update work items, track time, and manage custom fields. It is for teams using AI-powered IDEs or assistants who want to manage projects without switching contexts.
- Search and browse spaces, projects, and tasks from your Astravue workspace while working in an AI tool.
- Create, update, and comment on tasks and subtasks using natural language instructions in Claude or Cursor.
- Track time entries with live timers or manual logging and generate timesheet reports without leaving your IDE.
Tools this server exposes
| Tool | Does |
|---|---|
| search_spaces | Search and retrieve Astravue spaces. |
| search_projects | Search and retrieve Astravue projects. |
| search_tasks | Search and retrieve Astravue tasks. |
| create_task | Create a new task in Astravue. |
| update_task | Update an existing task in Astravue. |
| create_subtask | Create a subtask within an Astravue task. |
| create_comment | Add a comment to an Astravue task. |
| create_time_entry | Create a time entry or start a live timer for a task. |
| get_timesheet | Retrieve timesheet data and time tracking reports. |
| manage_custom_fields | Read and update custom fields including dropdowns, dates, and currencies. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.