datadabble-mcp

DataDabble MCP Server

A Model Context Protocol (MCP) server that wraps the DataDabble REST API, enabling Claude and other MCP clients to manage databases, fields, entries, visualizations, AI features, notifications, audit logs, and team accounts via natural language.

The server communicates with DataDabble's REST API rather than accessing MongoDB directly, respecting auth, validation, and business logic in the app layer.

Quick Start

Prerequisites

• Python 3.11+
• A running DataDabble instance (default: http://localhost:5000)

Install

pip install -e .

Configure

Copy .env.example to .env and set your values:

DATADABBLE_BASE_URL=http://localhost:5000/api/v1
DATADABBLE_EMAIL=user@example.com
DATADABBLE_PASSWORD=password
DATADABBLE_AUTO_LOGIN=true

Run

# As a module
python -m datadabble_mcp

# Or via the entry point
datadabble-mcp

Claude Desktop Configuration

Add to your Claude Desktop config (claude_desktop_config.json):

{
  "mcpServers": {
    "datadabble": {
      "command": "python",
      "args": ["-m", "datadabble_mcp"],
      "env": {
        "DATADABBLE_BASE_URL": "http://localhost:5000/api/v1",
        "DATADABBLE_EMAIL": "user@example.com",
        "DATADABBLE_PASSWORD": "password"
      }
    }
  }
}

Tools (42)

Auth (4)

| Tool | Description | |------|-------------| | auth_login | Log in with email and password | | auth_logout | Log out and clear tokens | | auth_whoami | Get current user profile | | auth_register | Register a new account |

Databases (5)

| Tool | Description | |------|-------------| | db_list | List all databases | | db_create | Create a new database | | db_get | Get a database by slug | | db_update | Update title or description | | db_delete | Delete a database |

Fields (7)

| Tool | Description | |------|-------------| | field_list | List fields in a database | | field_create | Add a field (BOOL, INT, DEC, STR, DATE, EMAIL, URL, DICT, LIST) | | field_get | Get a field by ID | | field_update | Update a field (with optional data loss confirmation) | | field_delete | Delete a field | | field_reorder | Reorder fields | | field_preview_type_change | Preview impact of changing a field's type |

Entries (6)

| Tool | Description | |------|-------------| | entry_list | List entries with pagination and filtering | | entry_create | Create an entry | | entry_get | Get an entry by ID | | entry_update | Update an entry's values | | entry_delete | Delete an entry | | entry_validate_filter | Validate a filter expression |

Visualizations (7)

| Tool | Description | |------|-------------| | viz_list | List saved visualizations | | viz_create | Create a visualization (bar, line, pie, scatter) | | viz_get | Get a visualization by ID | | viz_update | Update a visualization | | viz_delete | Delete a visualization | | viz_data | Get computed data for a visualization | | viz_adhoc_data | Get visualization data without saving |

AI (3)

| Tool | Description | |------|-------------| | ai_insights | Get AI-generated insights about a database | | ai_ask | Ask a natural language question about data | | ai_suggest_query | Get a suggested filter from a description |

Notifications (7)

| Tool | Description | |------|-------------| | notification_list | List notifications | | notification_unread_count | Get unread count | | notification_mark_read | Mark one as read | | notification_mark_all_read | Mark all as read | | notification_delete | Delete a notification | | notification_prefs_get | Get notification preferences | | notification_prefs_update | Update notification preferences |

Audit Logs (6)

| Tool | Description | |------|-------------| | audit_list | List audit logs for a database | | audit_stats | Get audit statistics for a database | | audit_account_list | List account-level audit logs | | audit_account_stats | Get account-level audit statistics | | audit_account_users | List users in audit logs | | audit_export | Export audit logs as CSV |

Accounts & Teams (7)

| Tool | Description | |------|-------------| | account_get | Get current account details | | account_list | List all accounts | | account_switch | Switch account context | | account_members | List team members | | account_invite | Invite a user by email | | account_update_member | Update a member's role/permissions | | account_remove_member | Remove a team member |

Resources

| URI | Description | |-----|-------------| | datadabble://field-types | Field type reference | | datadabble://filter-syntax | Filter expression language reference | | datadabble://databases | List all databases (live) | | datadabble://databases/{slug} | Database schema with fields and entry count | | datadabble://databases/{slug}/sample | First 10 entries for context |

Prompts

| Name | Description | |------|-------------| | create-database | Create a database and add fields from a spec | | analyze-database | Fetch schema, sample data, and AI insights | | query-builder | Build a filter expression from natural language | | bulk-create-entries | Create multiple entries from a JSON array |

Development

# Install with dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Lint
ruff check src/ tests/

Architecture

src/datadabble_mcp/
├── server.py          # FastMCP server, lifespan, entry point
├── config.py          # Environment variable configuration
├── auth.py            # JWT auth manager (login, refresh, logout)
├── client.py          # httpx client with auth injection + auto-refresh
├── resources.py       # MCP resources (field types, filter syntax, schemas)
├── prompts.py         # MCP prompts (workflow templates)
└── tools/
    ├── auth_tools.py
    ├── database_tools.py
    ├── field_tools.py
    ├── entry_tools.py
    ├── viz_tools.py
    ├── ai_tools.py
    ├── notification_tools.py
    ├── audit_tools.py
    └── account_tools.py

The server uses stdio transport and auto-logs in on startup when credentials are configured. All API errors are returned as structured dicts (never raised as exceptions), so MCP clients always get a useful response.