Connect AI tools to Tabple
Issue a personal connection token so external tools can use your Tabple data safely.
Issuing a connection token
Settings → MCP & API tokens
Issue a new token from the 'MCP & API tokens' card under your profile avatar > Settings. The token is shown in plain text only when it is issued; after that only a hash is stored.
Revoking a token
Use Revoke on the token card to invalidate it immediately. After revoking, no request with that token passes.
Claude Desktop / Cursor setup
Edit claude_desktop_config.json
Add the Tabple MCP server to the Claude Desktop config file. Path on macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"tabple": {
"command": "npx",
"args": ["mcp-remote", "https://tabple.com/api/mcp", "--header", "Authorization:${AUTH_HEADER}"],
"env": {
"AUTH_HEADER": "Bearer YOUR_PAT_HERE"
}
}
}
}Restart Claude
Restart Claude Desktop completely and the 'tabple' server appears in the tool list. Cursor and VS Code are configured the same way.
The 20 available tools
- list_projects — your project list
- list_columns — the board columns of a project (find the IDs used to create and move tasks)
- get_tasks / create_task / update_task — read, create, and update tasks
- list_pages / get_page — the wiki page tree and page bodies
- get_wiki_block_guide — reference for wiki block formats
- create_page / update_page / delete_page — create, update, and delete wiki pages (children included)
- get_project_stats — project stats (task counts, progress)
- search — search tasks, wiki pages, and feature specs in the projects you can access
- get_prd / update_prd — read and partially update the PRD (only the fields you supply change)
- list_features / create_feature / update_feature / delete_feature — feature spec CRUD (level 1-3 hierarchy)
- link_feature_task — link or unlink a feature spec item and a task (action=link|unlink)
Limits
Free limits
10 per minute / 60 per hour / 300 per day. Enough for a light personal setup.
Paid plan limits
Starter, Pro, and Lifetime get 30 per minute / 300 per hour / 5,000 per day. Built for heavy agent workflows.
When you go over
You get a 429 response with a Retry-After header. AI agents back off automatically.