Astravue MCP Server
The Astravue MCP Server is a cloud-based bridge between your Astravue workspace and compatible AI tools. Once configured, it enables those tools to interact with your spaces, projects, tasks, time entries, and custom fields in real time. All actions are authenticated via OAuth 2.0 and respect the user's existing access controls.
With the Astravue MCP Server, you can:
- Search and browse spaces, projects, and tasks without switching tools.
- Create and update tasks, subtasks, checklists, and comments through natural language.
- Track time with live timers or manual entries and generate timesheet reports.
- Manage custom fields including dropdowns, dates, currencies, and more.
- Automate workflows like sprint planning, status reporting, and bulk task updates.
It's designed for project managers, developers, and teams who use AI-powered IDEs or assistants and want to manage work without context switching.
Full product documentation lives at docs.astravue.com.
Astravue is a verified and reviewed integration in the Claude Connectors Directory.
Supported Clients
All clients connect to the same endpoint: https://api.astravue.com/mcp
You will be asked to sign in to Astravue in your browser the first time you connect.
Setup Instructions
Claude Connectors Directory (recommended)
Astravue is available as a verified connector across Claude.ai, Claude Desktop, Claude Mobile, Claude Code, and Cowork.
- In Claude.ai or Claude Desktop, go to Customize → Connectors.
- Click +, browse the Connectors Directory, and search for Astravue.
- Select Astravue, click Connect, then sign in to Astravue and approve access.
- Start a new conversation and enable Astravue from the + → Connectors menu.
When Claude Code is authenticated with your Claude.ai subscription, the connector is available automatically. Run /mcp to confirm it is connected. Claude Code sessions using an API key, Bedrock, or Vertex do not load Claude.ai connectors; use the direct connection command below instead.
Claude Code
The Connectors Directory provides the verified MCP connection. For eight guided project-management workflows as well, install the Astravue plugin. A direct MCP connection remains available for API-key, Bedrock, Vertex, and custom setups.
Option A — Install as a Claude Code plugin (recommended)
The plugin bundles the MCP server with eight pre-built skills that automate common project management workflows. It is listed in Anthropic's community plugin marketplace.
Inside Claude Code, run:
/plugin marketplace add anthropics/claude-plugins-community
/plugin install astravue@claude-community
/reload-plugins
Alternatively, install directly from this repository:
/plugin marketplace add AstravueOrg/astravue-mcp-server
/plugin install astravue@astravue
/reload-plugins
After install, run /mcp to confirm astravue is connected, then open /plugin and switch to the Installed tab to view the bundled skills.
See Anthropic's plugin discovery docs for installation scopes (user, project, local).
Option B — Install only the MCP server
If you only want the raw MCP server without the bundled skills, run:
claude mcp add --transport http astravue https://api.astravue.com/mcp
Your browser will open automatically. Sign in with your Astravue account and click Approve.
To confirm the connection, run /mcp in a new Claude Code session. You should see astravue listed with a green connected status.
Add the --scope project flag to create a .mcp.json file in your project root:
claude mcp add --transport http --scope project astravue https://api.astravue.com/mcp
Commit .mcp.json to your repository so teammates automatically get the connection.
Claude Desktop
- Go to Customize → Connectors in Claude Desktop.
- Click +, browse the directory, and search for Astravue.
- Select Astravue, click Connect, then sign in and approve access.
If Connectors is not available in your Claude Desktop version, edit:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"astravue-mcp": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}
Save the file and restart Claude Desktop.
Claude.ai (Web)
- Go to claude.ai → Customize → Connectors.
- Click +, browse the directory, and search for Astravue.
- Select Astravue, click Connect, then sign in and approve access.
Claude Mobile (iOS / Android)
Claude Mobile uses the same connector ecosystem as Claude.ai. Connect Astravue from the mobile Connectors Directory, or connect it once on Claude.ai or Claude Desktop and use the same Claude account on mobile.
Cursor
- Go to Cursor Settings → Tools & MCP
- Click New MCP Server — this opens your
mcp.jsonfile - Add the following configuration:
{
"mcpServers": {
"astravue-mcp": {
"url": "https://api.astravue.com/mcp"
}
}
}
- Save the file. Your browser will open — sign in with your Astravue account and click Approve.
Create a .cursor/mcp.json file in your project root with the same config above. This scopes the connection to that project only.
VS Code
- Add a
.vscode/mcp.jsonfile to your project root:
{
"servers": {
"astravue-mcp": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}
- Open Command Palette (
Cmd+Shift+Pon macOS,Ctrl+Shift+Pon Windows), run MCP: List Servers, and click Start next toastravue-mcp. - Authenticate with your Astravue account in the browser.
Add this to your User Settings (JSON):
{
"mcp": {
"servers": {
"astravue-mcp": {
"type": "http",
"url": "https://api.astravue.com/mcp"
}
}
}
}
Windsurf
-
Open the Windsurf MCP config file:
- macOS / Linux:
~/.codeium/mcp_config.json - Windows:
C:\Users\[username]\.codeium\mcp_config.json - Or go to Settings → Tools → Windsurf Settings → Add Server
- macOS / Linux:
-
Add the following:
{
"mcpServers": {
"astravue-mcp": {
"serverUrl": "https://api.astravue.com/mcp"
}
}
}
If you already have other servers in your config, add
astravue-mcpalongside them inside the existingmcpServersobject — don't replace the entire file.
- Reload Windsurf and authenticate with your Astravue account.
Cline
- Click the Cline icon in the VS Code sidebar → click the menu (⋮) in the top right → select MCP Servers.
- Go to the Remote Servers tab and fill in:
- Server Name:
astravue-mcp - Server URL:
https://api.astravue.com/mcp - Transport Type:
Streamable HTTP
- Server Name:
- Click Add Server. Your browser will open — sign in and click Approve.
Click Configure MCP Servers in the Configure tab to open cline_mcp_settings.json and add:
{
"mcpServers": {
"astravue-mcp": {
"url": "https://api.astravue.com/mcp",
"type": "streamableHttp"
}
}
}
JetBrains (IntelliJ, WebStorm, PyCharm, etc.)
- Go to Settings → Tools → AI Assistant → Model Context Protocol (MCP)
- Click Add → Select Streamable HTTP → Paste:
{
"mcpServers": {
"astravue-mcp": {
"url": "https://api.astravue.com/mcp"
}
}
}
- Click OK, then Apply. Sign in with your Astravue account in the browser.
Gemini CLI
Copy gemini-extension.json from this repository to your Gemini CLI extensions directory, or add the server URL to your Gemini MCP configuration.
Other MCP Clients
For any client that supports Streamable HTTP + OAuth 2.0:
If your client only supports SSE transport, run:
npx -y mcp-remote@latest https://api.astravue.com/mcp
Point your client to the local mcp-remote process instead of the Astravue URL.
How It Works
Astravue MCP exposes workspace operations as tools that AI assistants can call. Each tool performs a specific action in your Astravue workspace such as creating projects, managing tasks, logging time, or retrieving reports.
When an AI assistant receives a request, it selects the appropriate tool based on the user's prompt and executes it with the required parameters.
Authentication: All tools require an active authenticated session. If a tool call fails with
401 Unauthorized, reconnect your AI client to refresh your session.
Destructive actions: Some tools perform destructive actions such as deleting tasks, projects, or spaces. Most AI clients require confirmation before executing these operations. Always review actions before approving them.
Permissions
Astravue MCP respects the same role-based access control as the Astravue web app. Every tool call is executed as your authenticated user with your existing permissions. You cannot access or modify anything through MCP that you cannot access in the app.
Organization Roles
Space-Level Permissions
Project-Level Permissions
Task-Level Permissions
If a tool call returns
403 Access Denied, it means your account does not have the required role for that action. Contact your workspace admin to update your permissions.
Example Workflow
A user asks an AI assistant:
"Create a new sprint called Sprint 5 and add tasks for login, payments, and API docs."
The AI assistant may execute these tools:
astravue_create_project— creates the Sprint 5 projectastravue_bulk_create_tasks— creates the three tasks in one callastravue_update_task— sets priorities or due dates as needed
This allows complex project setup to be completed from a single conversation.
Available Tools
The Astravue MCP Server provides 87 tools across the following categories.
Available Prompts
The Astravue MCP Server provides 15 prompts — pre-built workflows that guide your AI assistant through common project management tasks.
Available Resources
The Astravue MCP Server provides 6 resources — context data that AI assistants can attach to conversations for richer, more accurate responses.
Static Resources
Resource Templates
Skills
This repository includes ready-to-use skills that automate common project management workflows:
Skills are available automatically when using the Claude Code or Cursor plugin.
Data and Security
- All traffic is encrypted via HTTPS using TLS 1.2 or later.
- OAuth 2.0 authentication ensures secure access control.
- Data access respects your workspace role and project-level permissions.
- The MCP server never stores credentials — authentication is handled per session.
For more details, see SECURITY.md.
Rate Limits
Astravue MCP enforces rate limits to ensure platform stability.
If the limit is exceeded, the server returns HTTP 429 Too Many Requests. Wait 60 seconds before retrying.
Troubleshooting
Sign out of Astravue at app.astravue.com, sign back in, then reconnect your AI client.
Your OAuth session expired. Reconnect your client to trigger a new sign-in flow.
Ensure your client supports Streamable HTTP transport. Clients that only support SSE must use the mcp-remote bridge (see Other MCP Clients).
OAuth tokens expire periodically. Your client should refresh automatically. If it does not, reconnect once.
FAQ
No. Astravue MCP works on all plans including free.
Yes. Each user authenticates with their own Astravue account.
Astravue MCP only has access to resources your account already has permission to access. All actions are performed as your authenticated user.
No. OAuth 2.0 is required so every action is tied to an authenticated user.
Most AI clients ask for confirmation before executing write operations. Check your client's tool approval settings.
Contributing
We welcome contributions. See CONTRIBUTING.md for guidelines.
Code of Conduct
This project follows the Contributor Covenant.
License
This project is licensed under the Apache License 2.0.
Copyright 2026 Astravue.
Support
- Open an issue on GitHub
- Email [email protected]
- Visit help.astravue.com