Dooist

Provides task-contextful MCP capabilities for Claude Code, storing rich task data and enabling seamless handoffs across sessions.

javascript

0

GitHub Stars

javascript

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "papermoose-dooist": {
      "command": "npx",
      "args": [
        "dooist"
      ],
      "env": {
        "DOOIST_DB_PATH": "~/.dooist/dooist.db"
      }
    }
  }
}

Dooist is an MCP server that stores rich task context for AI agents, enabling seamless, stateful task management across sessions. It lets agents pick up exactly where they left off by carrying intent, plan, status, and outputs inside each task, so you no longer need to re-describe goals or recreate context between Claude Code sessions.

How to use

You interact with Dooist through an MCP client integrated with your Claude Code workflow. Create tasks with full context, let the agent prioritize and assign work, and then continue from the same point in subsequent sessions. Use Dooist to keep track of what your agents should do next, what’s in progress, and what’s completed.

How to install

Prerequisites: ensure you have access to Claude Code with MCP support and a working Node.js environment if you plan to run local commands as shown.

claude mcp add dooist -- npx dooist

Restart Claude Code after adding the MCP server.

## Configuration and usage notes

You can customize where Dooist stores its data by providing an environment variable when adding the MCP server. For example, you can set the database path to a specific location so your data remains in a desired directory.

claude mcp add dooist -e DOOIST_DB_PATH=/your/path/dooist.db -- npx dooist

## Notes on data and hosting

Dooist is self-hosted, storing data locally on your machine. The persistent SQLite database lives by default at ~/.dooist/. This ensures your task context remains available across sessions without relying on remote services.

## Security and data exposure

Treat the Dooist store as you would any local data store. Control access to Claude Code and the machine where Dooist runs. Use standard security practices for your environment, including file permissions and secure access to the host where the MCP server operates.

## Troubleshooting tips

If tasks do not appear across sessions, verify that the MCP server is running, the database path is accessible, and your Claude Code session has the Dooist MCP configured. Check that the database file exists at the configured path and that Dooist has permission to read and write to it.

## Available tools

### create\_task

Add a new task with full context, including due date, priority, and labels.

### get\_task

 Retrieve the complete task details with intent, plan, and outputs.

### list\_tasks

List tasks with optional filters by project, label, or status.

### complete\_task

Mark a task as finished.

### update\_task

Modify any field of a task, including its context.

### delete\_task

Remove a task from the list.

### today

Show tasks due today and overdue items.

### upcoming

Show tasks due in the next N days.

### scan\_todos

Scan codebases for TODO/FIXME/HACK comments and create tasks from them.

### list\_projects

List all projects to which tasks can belong.

### create\_project

Create a new project to group related tasks.

### list\_labels

List all labels you use to categorize tasks.

### create\_label

Create a new label for task categorization.