Kommo CRM

io.github.depper-IAv1.0.3Updated Oct 3, 2026

Manage Kommo CRM leads, contacts, pipelines, tasks, notes, tags and companies.

VerifiedSTDIODesktop onlyBusiness & Commerce Productivity & Workflow Communication & Collaboration

Overview

AI-generated overview

Lets an assistant manage Kommo CRM leads, contacts, pipelines, tasks, notes, tags and companies through natural language.

What it does: Connects an AI assistant to a Kommo CRM account (formerly amoCRM) over the Kommo API v4. Tools cover leads (list, get, create, update, delete, move stage, bulk update), contacts, pipelines and stages, tags, tasks with an overdue filter, notes, chat templates and messages, companies, and custom fields. An advanced tool can create a lead, contact and company in one call.
When to use it: Useful when you work in Kommo CRM and want an assistant to look up or change records without opening the web interface: reviewing pipeline leads, creating contacts, moving deals between stages, or listing overdue tasks. It is aimed at desktop MCP clients such as Kiro, Claude Desktop, OpenCode, Cursor or Codex.
Requirements: Runs locally as a Python process (Python 3.10-3.13), typically via uvx from the PyPI package kommo-kiro-power. Needs a Kommo integration and OAuth credentials supplied by environment variable: KOMMO_SUBDOMAIN and KOMMO_ACCESS_TOKEN are required, with KOMMO_CLIENT_ID, KOMMO_CLIENT_SECRET, KOMMO_REDIRECT_URI and KOMMO_REFRESH_TOKEN used for token refresh. Network access to the Kommo API is required.

Before you install

The server holds Kommo OAuth credentials (KOMMO_ACCESS_TOKEN, KOMMO_CLIENT_SECRET, KOMMO_REFRESH_TOKEN) that grant access to your CRM data; keep them out of version control. Many tools write to the CRM, including create, update, delete, bulk update and stage moves, and the chat tool can send messages to customers, so changes may be hard to undo. The README states the server runs locally, talks only to your Kommo account, and sends no telemetry.

Installation

In SourceWeft

Open Kommo CRM in the dashboard and add it to a workspace.
Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

kommo-kiro-power — Kommo CRM Power for Kiro

[Kiro Power] [MCP Compatible] [Python 3.10+] [MIT License]

[Kommo Kiro MCP server - quality and maintenance score on Glama]

Connect AI agents to your Kommo CRM (formerly amoCRM) using the Model Context Protocol. Manage leads, contacts, pipelines, tasks, notes, tags, and companies — all from natural language in Kiro or any MCP-compatible client.

Features

Category	Tools
Leads	list, get, create, update, delete, move stage, bulk update
Contacts	list, get, create, update
Pipelines	list, create, update
Stages	list, create, update
Tags	list, add, remove (auto-creates if needed)
Tasks	create, list (with overdue filter)
Notes	add note to lead
Chat	list templates, send message
Companies	create, list
Custom Fields	list, create (text, select, multiselect, date, etc.)
Advanced	create lead + contact + company in one call

Quick Start

Install as Kiro Power

Open Kiro → Powers panel → Add Custom Power
Select Import power from GitHub
Paste: https://github.com/depper-IA/kommo-kiro-power
Follow the onboarding steps when activated

Install as standalone MCP

Published on PyPI and the official MCP Registry. With uv installed, no clone is needed:

Or install the one-click MCPB bundle (kommo-crm.mcpb) in Claude Desktop or any MCPB-compatible client. Bundle sources live in mcpb/. After changing tools, run python scripts/build_mcpb.py to sync the manifest, then pnpm dlx @anthropic-ai/mcpb pack mcpb kommo-crm.mcpb. For Smithery, python scripts/build_mcpb.py --smithery kommo-crm-smithery.mcpb builds a variant that also carries each tool's inputSchema.

Or from source:

Setup

1. Create a Kommo Integration

Go to Kommo → Settings → Integrations → Create integration
Copy the Client ID and Client Secret
Set http://localhost:8080/callback as the Redirect URI

2. Configure Credentials

3. Authenticate (OAuth)

A browser opens → sign in to Kommo → authorize → tokens saved automatically.

4. Run the MCP Server

Usage with Kiro

Once installed as a Power, the server activates automatically when you mention keywords like "kommo", "leads", "pipeline", "crm", or "contacts" in your conversation.

"Show me all leads in the Sales pipeline""Create a lead for Maria from TechCorp with phone +57 310 543 6281""Move lead 12345 to the Negotiation stage""List all overdue tasks"

"Show me all leads in the Sales pipeline""Create a lead for Maria from TechCorp with phone +57 310 543 6281""Move lead 12345 to the Negotiation stage""List all overdue tasks"

Usage with Other MCP Clients

Add to your claude_desktop_config.json, opencode.json, or equivalent:

json

{  "mcpServers": {    "kommo": {      "command": "uvx",      "args": ["--from", "kommo-kiro-power", "kommo-mcp"],      "env": {        "KOMMO_SUBDOMAIN": "your_subdomain",        "KOMMO_ACCESS_TOKEN": "your_access_token"      }    }  }}

{  "mcpServers": {    "kommo": {      "command": "uvx",      "args": ["--from", "kommo-kiro-power", "kommo-mcp"],      "env": {        "KOMMO_SUBDOMAIN": "your_subdomain",        "KOMMO_ACCESS_TOKEN": "your_access_token"      }    }  }}

Architecture

kommo-kiro-power/├── plugin.json           # Kiro Power manifest (Agent Plugins format)├── POWER.md              # Legacy Kiro Power metadata + onboarding├── mcp.json              # MCP server config for Kiro├── server.json           # Official MCP Registry metadata├── skills/               # Workflow guides loaded on-demand│   ├── leads-workflow/SKILL.md│   ├── pipeline-management/SKILL.md│   └── automation-patterns/SKILL.md├── kommo_mcp/            # Python MCP server│   ├── __main__.py       # Entry point (stdio transport)│   ├── mcp_server.py     # Server + tool registration│   ├── kommo_client.py   # HTTP client (OAuth, retry, cache)│   └── tools/            # Tool definitions + handlers│       ├── leads.py│       ├── contacts.py│       └── pipelines.py├── scripts/│   └── oauth_setup.py    # One-time OAuth authentication├── pyproject.toml        # Package config└── requirements.txt

kommo-kiro-power/├── plugin.json           # Kiro Power manifest (Agent Plugins format)├── POWER.md              # Legacy Kiro Power metadata + onboarding├── mcp.json              # MCP server config for Kiro├── server.json           # Official MCP Registry metadata├── skills/               # Workflow guides loaded on-demand│   ├── leads-workflow/SKILL.md│   ├── pipeline-management/SKILL.md│   └── automation-patterns/SKILL.md├── kommo_mcp/            # Python MCP server│   ├── __main__.py       # Entry point (stdio transport)│   ├── mcp_server.py     # Server + tool registration│   ├── kommo_client.py   # HTTP client (OAuth, retry, cache)│   └── tools/            # Tool definitions + handlers│       ├── leads.py│       ├── contacts.py│       └── pipelines.py├── scripts/│   └── oauth_setup.py    # One-time OAuth authentication├── pyproject.toml        # Package config└── requirements.txt

Security

Credentials read from .env — never hardcoded
.env is in .gitignore — never committed
OAuth v4 with automatic token refresh
Rate limiting: max 5 requests/second
Exponential backoff with jitter on failures
Response caching for pipelines/stages/fields (10 min TTL)

Compatibility

Python: 3.10, 3.11, 3.12, 3.13
MCP Clients: Kiro, Claude Desktop, OpenCode, Cursor, Codex
Kommo API: v4

Privacy

The server runs locally on your machine and talks only to your own Kommo account API. It has no telemetry and sends no data to the author or any third party. See PRIVACY.md.

Support

Bugs and feature requests: GitHub Issues
Email: [email protected]

License

MIT

Source: README.md at commit 593395f

Tools

Tool metadata has not been indexed yet.

Version history

v1.0.3LatestOct 3, 2026