Coherence

Name: Coherence
Author: getcoherence

@getcoherence

关于 Coherence

MCP (Model Context Protocol) server that exposes Coherence as tools for Claude, Cursor, Cline, and other MCP-compatible clients.

基本信息

分类

其他

许可证

MIT

运行时

node

传输方式

stdio

发布者

getcoherence

提交者

Keith

配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "coherence": {
      "command": "npx",
      "args": [
        "-y",
        "@coherenceos/mcp-server"
      ],
      "env": {
        "COHERENCE_API_URL": "https://api.getcoherence.io/v1",
        "COHERENCE_API_TOKEN": "<YOUR_API_KEY>"
      }
    }
  }
}

工具

Send a message to a Coherence agent; the agent uses its own toolset (email, reminders, calendar, social, landing pages, approvals, etc.) and returns its response

List CRM modules in the workspace

List fields on a module — useful before creating/updating records

List records in a module with search + pagination

Get a single record by ID

Create a new record

Update a record's fields

Soft-delete a record (recoverable from trash)

List blog articles, LinkedIn posts, emails, etc.

Get a single outreach item, including its MDX content

Create a new outreach item

Update the MDX body of an outreach item

Update outreach metadata (title, status, etc.)

概览

What is Coherence?

Coherence is an MCP (Model Context Protocol) server that exposes the Coherence workspace platform as tools for Claude, Cursor, Cline, and other MCP-compatible clients. Its headline tool, chat_with_agent, lets you send a message to your Coherence agent (Nash) to draft emails, create reminders, update records, post to social media, create landing pages, and more—all under your workspace’s approval rules.

How to use Coherence?

Create an API key in Coherence (Settings → API Keys → Create new key), then add the provided JSON snippet to your MCP client’s configuration file (e.g., claude_desktop_config.json for Claude Desktop). Restart the client and start talking to your workspace using natural language commands like "Ask my Coherence agent to draft a follow-up email to the leads I created this week."

Key features of Coherence

chat_with_agent sends messages to a full-toolset Coherence agent
Direct CRUD operations on module records (create, read, update, delete)
List and search records with pagination
Manage outreach items (blog articles, LinkedIn posts, emails, etc.)
View and update record fields with list_module_fields
All operations respect workspace approval gates

Use cases of Coherence

Draft and send follow-up emails to leads created in the current week
List open deals over a specified dollar amount from the CRM
Create a new contact and set a reminder for follow-up
Manage landing pages, calendar events, and social media posts via the agent

FAQ from Coherence

What runtime does Coherence require?

The server runs via npx (Node.js) and is installed as a package (@coherenceos/mcp-server). No additional local setup is needed beyond a Node.js environment.

How is authentication handled?

You must provide a COHERENCE_API_TOKEN (an sk_live_… key) from your Coherence workspace. The key can be scoped to specific permissions (records, collab, agents, workspace) or inherit your full user permissions.

What data does the server access?

All data lives in your Coherence workspace (CRM records, outreach items, etc.). The server only interacts with the Coherence API; no data is stored locally.

Are there any built-in security limits?

Approval gates are enforced inside the Coherence platform regardless of API key scopes. An API key cannot bypass workspace-level approval rules for actions like sending email or posting to social media.

Which transport does the MCP server use?

The server uses the standard stdio transport (command-line invocation via npx). It is configured as a subprocess in the MCP client’s JSON configuration.

其他分类下的更多 MCP 服务器

AutoBrowser MCP

autobrowser-ai

Browser MCP is a Model Context Provider (MCP) server that allows AI applications to control your browser

6.8K

Activepieces

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

23.2K