Daeda MCP

An MCP server that syncs your HubSpot CRM to a local encrypted database, enabling AI assistants to query your contacts, companies, and deals instantly.

Why Daeda?

Querying HubSpot through the API is slow and rate-limited. Daeda solves this by:

• Syncing your entire CRM locally - Contacts, companies, deals, and all associations
• Encrypted storage - Your data is encrypted at rest using your HubSpot token
• Instant queries - AI assistants can run SQL queries against your local database
• Works offline - Once synced, no internet required for queries

Features

• Full CRM Sync - Exports all contacts, companies, and deals with all properties
• Association Support - Contact-company, deal-contact, and deal-company relationships
• Smart Seeding - Quick preview of ~1,000 recent deals available in seconds while full sync runs
• Resumable Sync - Interrupted syncs resume automatically on restart
• Read-Only Queries - AI can only SELECT data, never modify your CRM

Use Cases

• "Show me all deals closing this month over $50k"
• "Find contacts at companies in the healthcare industry"
• "Which deals have no associated contacts?"
• "List all contacts with a @gmail.com email"
• "What's the total pipeline value by deal stage?"

Prerequisites

HubSpot Private App Setup

1. Go to your HubSpot account → Settings → Integrations → Private Apps
2. Create a new private app
3. Under "Scopes", enable these permissions:
   • crm.export (required for bulk data export)
   • crm.objects.contacts.read
   • crm.objects.companies.read
   • crm.objects.deals.read
4. Create the app and copy your access token (starts with pat-)

Quick Start

No installation needed! Just add Daeda to your AI assistant's MCP configuration below - npx handles everything automatically.

Configuration

Choose your AI assistant below and follow the setup instructions. All configurations use the same JSON structure - just add it to your tool's config file.

| Platform | Config File Location | |----------|---------------------| | macOS | ~/Library/Application Support/Claude/claude_desktop_config.json | | Windows | %APPDATA%\Claude\claude_desktop_config.json |

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/.cursor/mcp.json or .cursor/mcp.json in project | | Windows | %USERPROFILE%\.cursor\mcp.json or .cursor\mcp.json in project |

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/.codeium/windsurf/mcp_config.json | | Windows | %USERPROFILE%\.codeium\windsurf\mcp_config.json |

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/Library/Application Support/Code/User/settings.json | | Windows | %APPDATA%\Code\User\settings.json |

Add to your VS Code settings.json:

{
  "mcp": {
    "servers": {
      "daeda": {
        "command": "npx",
        "args": ["-y", "daeda-mcp"],
        "env": {
          "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
        }
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json | | Windows | %APPDATA%\Code\User\globalStorage\saoudrizwan.claude-dev\settings\cline_mcp_settings.json |

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/Library/Application Support/Code/User/globalStorage/rooveterinaryinc.roo-cline/settings/mcp_settings.json | | Windows | %APPDATA%\Code\User\globalStorage\rooveterinaryinc.roo-cline\settings\mcp_settings.json |

You can also edit via the Roo Code UI: Click the MCP icon → Edit Global MCP.

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

Windows users: Use this format instead:

{
  "mcpServers": {
    "daeda": {
      "command": "cmd",
      "args": ["/c", "npx", "-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/.config/zed/settings.json | | Windows | %APPDATA%\Zed\settings.json |

Add to the context_servers section of your Zed settings:

{
  "context_servers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

JetBrains IDEs (2025.2+) have built-in MCP support via Settings → Tools → AI Assistant → Model Context Protocol (MCP).

1. Open Settings → Tools → AI Assistant → Model Context Protocol (MCP)
2. Click + to add a new server
3. Paste this configuration:

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

Config is stored internally at: | Platform | Location | |----------|----------| | macOS | ~/Library/Application Support/JetBrains/<IDE>/options/mcp_servers.xml | | Windows | %APPDATA%\JetBrains\<IDE>\options\mcp_servers.xml |

| Platform | Config File Location | |----------|---------------------| | macOS | ~/Library/Application Support/amazon-q/mcp.json | | Windows | %APPDATA%\amazon-q\mcp.json | | Linux | ~/.config/amazon-q/mcp.json |

Create the file if it doesn't exist:

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

Restart your IDE after saving.

| Platform | Config File Location | |----------|---------------------| | macOS | ~/.claude/config.json (or inherits from Claude Desktop) | | Windows | %USERPROFILE%\.claude\config.json |

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/.continue/config.yaml | | Windows | %USERPROFILE%\.continue\config.yaml |

Continue uses YAML format:

mcpServers:
  - name: daeda
    command: npx
    args:
      - "-y"
      - "daeda-mcp"
    env:
      HS_PRIVATE_TOKEN: "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"

| Platform | Config File Location | |----------|---------------------| | macOS | ~/.config/goose/config.yaml | | Windows | %APPDATA%\Block\goose\config\config.yaml |

Goose uses YAML format:

extensions:
  daeda:
    name: daeda
    type: stdio
    cmd: npx
    args: ["-y", "daeda-mcp"]
    envs:
      HS_PRIVATE_TOKEN: "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"

Raycast supports MCP via the "Model Context Protocol" extension.

1. Install the MCP extension from Raycast Store
2. Open Raycast → Extensions → Model Context Protocol → Settings
3. Import this configuration:

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | Global | ~/.config/opencode/opencode.json | | Project | <PROJECT_ROOT>/opencode.json |

{
  "mcpServers": {
    "daeda": {
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

| Platform | Config File Location | |----------|---------------------| | macOS | ~/Library/Application Support/Trae/User/settings.json | | Windows | %APPDATA%\Trae\User\settings.json |

Add to your Trae settings:

{
  "trae.mcp.servers": {
    "daeda": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "daeda-mcp"],
      "env": {
        "HS_PRIVATE_TOKEN": "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
      }
    }
  }
}

Add to your librechat.yaml configuration:

mcpServers:
  daeda:
    type: stdio
    command: npx
    args: ["-y", "daeda-mcp"]
    env:
      HS_PRIVATE_TOKEN: "pat-xx-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"

How It Works

1. First Launch - Daeda requests bulk exports from HubSpot's Export API
2. Quick Seed - While exports process, ~1,000 recent deals are fetched via Search API for immediate use
3. Full Sync - Export CSVs are downloaded and streamed into an encrypted SQLite database
4. Ready - AI assistants can now query your full CRM instantly

The quick seed completes in 2-5 minutes, giving you immediate access to your 1,000 most recent deals and their associated contacts and companies.

The full sync runs in the background and duration depends on your CRM size:

• Small CRM (10k records): ~5-10 minutes
• Medium CRM (100k records): ~30-60 minutes
• Large CRM (1M+ records): up to 5 hours

Progress is shown via the db_status tool. You can start querying immediately after the quick seed completes.

Available Tools

| Tool | Description | |------|-------------| | db_status | Check sync progress and database health | | get_raw_sql | Execute SELECT queries against your CRM data |

Database Schema

The local database contains:

• contacts - All contacts with email and full properties as JSON
• companies - All companies with domain and full properties as JSON
• deals - All deals with name and full properties as JSON
• contact_company - Contact to company associations
• deal_contact - Deal to contact associations
• deal_company - Deal to company associations

Query any HubSpot property using json_extract():

SELECT 
  json_extract(properties, '$.firstname') as first_name,
  json_extract(properties, '$.lastname') as last_name,
  email
FROM contacts
WHERE json_extract(properties, '$.lifecyclestage') = 'customer'
LIMIT 10

Data Storage

Your CRM data is stored locally at:

• macOS/Linux: ~/.daeda-mcp/data/
• Windows: %APPDATA%\daeda-mcp\data\

The database is encrypted using your HubSpot token as the encryption key. If you change tokens, the database will be re-initialized automatically.

Security

• All data stays on your machine
• Database is encrypted at rest
• Only SELECT queries are allowed
• Dangerous SQL keywords are blocked
• Your HubSpot token is never stored (only used for encryption)

License

MIT - see LICENSE for details.