Skip to main content
Glama
Leonchen3352

SiliMind MCP Server

by Leonchen3352

SiliMind MCP Server

Connect Claude Desktop / ChatGPT / Cursor (any MCP-compatible client) to your SiliMind CRM.

Ask your AI assistant things like:

  • "Show me my open deals this week"

  • "What's on my calendar tomorrow?"

  • "Create a follow-up task for Acme Corp"

Your AI talks to SiliMind through read-only (and opt-in write-proposal) MCP tools — with your per-user permissions, tenant isolation, and full audit trail.

What it does

SiliMind exposes an external MCP gateway at /mcp/external (Streamable HTTP, JSON-RPC 2.0). This package provides:

  1. Python client library (silimind_mcp.SilimindMCPClient) — for building custom integrations

  2. CLI (python -m silimind_mcp.server) — list tools, call a tool from terminal

  3. Stdio bridge (python -m silimind_mcp.stdio_bridge) — for clients that only speak stdio (Claude Desktop, Cursor)

Related MCP server: namakan-mcp-crm

Available tools (read-only)

Tool

Description

search_customers

Search customers (respects your permissions)

get_customer_detail

Get full customer profile

list_deals / get_deal

Browse opportunities

dashboard_summary

Your business overview

query_data

Ask questions in natural language (SELECT-only, whitelisted tables)

list_contracts / get_contract

View contracts

list_tasks

Your task list

list_calendar

Merged calendar view

list_outbound_jobs

Outbound prospecting progress

Write-proposal tools (opt-in, readwrite scope tokens only)

Write tools never write directly — they create a proposal that you approve inside SiliMind:

Tool

Description

crm_create_contact

Propose creating a new contact

crm_create_deal

Propose creating a new deal

create_task

Propose creating a task

crm_add_interaction

Propose adding a timeline interaction

Proposals expire after 24h if not approved. Nothing touches your CRM without your explicit approval.

Setup

1. Get your MCP token

In SiliMind: Settings → MCP Access → Generate token (readonly or readwrite scope).

2. Configure your client

Claude Desktop (~/.claude/claude_desktop_config.json):

{
  "mcpServers": {
    "silimind": {
      "command": "python",
      "args": ["-m", "silimind_mcp.stdio_bridge"],
      "env": {
        "SILIMIND_URL": "https://agent.silimind.net",
        "SILIMIND_TOKEN": "smcp_your_token_here"
      }
    }
  }
}

Any HTTP MCP client: point it at https://agent.silimind.net/mcp/external with header Authorization: Bearer smcp_....

CLI quick test:

pip install silimind-mcp
export SILIMIND_URL=https://agent.silimind.net
export SILIMIND_TOKEN=smcp_...
python -m silimind_mcp.server list

Security

  • Per-token permissions (readonly / readwrite scope)

  • Tenant isolation & row-level security enforced server-side

  • Rate limiting (per-token per-minute + daily quota)

  • Every call audit-logged

  • Write operations require human approval inside SiliMind

Install

pip install silimind-mcp

Or from source:

git clone https://github.com/Leonchen3352/silimind-mcp.git
cd silimind-mcp
pip install -e .

License

Apache-2.0

Related MCP Connectors

Related MCP Servers

  • A
    license
    B
    quality
    C
    maintenance
    Enables AI assistants to manage RogerRoger CRM data including people, organizations, lists, tags, and tasks through standardized MCP tools.
    22
    6 npm
    MIT
  • A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI agents to interact with multiple CRM systems through a unified set of MCP tools, such as finding contacts, accounts, and deals, while supporting mock, REST, and vendor-specific backends.
    MIT
  • A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI agents to read and write CRM data—companies, contacts, opportunities, tasks, and interactions—through MCP, with OAuth and provenance tracking for every value.
    MIT
  • A
    license
    A
    quality
    B
    maintenance
    Enables AI agents to perform go-to-market workflows such as company and contact enrichment, CRM querying, and controlled, audited writes to a CRM through MCP tools callable from any MCP client.
    1
    MIT