Skip to main content

MCP server for Arythmatic Connect — let your AI agent post, comment, read channels, file support tickets, and chat from inside Claude / Cursor / any MCP-compatible client.

Project description

Arythmatic Connect — MCP Server

A Model Context Protocol server that lets your AI agent (Claude Desktop, Cursor, Continue, or any MCP-compatible host) post, comment, read channels, file support tickets, and chat inside an Arythmatic Connect workspace.

Install

pip install arythmatic-connect-mcp

or run from source:

cd mcp/
pip install -e .

Configure

The server authenticates as the tenant API key owner — every action the agent takes is performed as if that user did it (so the access scopes match what they normally have in the community UI). Mint a key at https://<your-tenant>.connect.arythmatic.cloud/admin/api-keys and copy the arc_live_<prefix>_<secret> value.

Set two env vars:

export ARYTHMATIC_API_BASE_URL="https://api.connect.arythmatic.cloud/api/v1"
export ARYTHMATIC_API_KEY="arc_live_xxxxxxxx_yyyyyyyyyyyy"

Claude Desktop wiring

In ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "arythmatic": {
      "command": "arythmatic-connect-mcp",
      "env": {
        "ARYTHMATIC_API_BASE_URL": "https://api.connect.arythmatic.cloud/api/v1",
        "ARYTHMATIC_API_KEY": "arc_live_xxxxxxxx_yyyyyyyyyyyy"
      }
    }
  }
}

Restart Claude Desktop. The Connect tools appear in the tools picker.

Cursor / Continue wiring

Cursor and Continue follow the same mcpServers config block — refer to their docs for the file location.

Tools

Discovery

Tool Description
get_current_user Who the agent is acting as + their permission scopes.
list_workspaces Workspaces the caller is a member of.
list_channels Channels inside a workspace.
search_content Cross-tenant search across posts, articles, channel messages.

Channel chat

Tool Description
list_channel_messages Recent messages in a channel (paginated).
post_channel_message Send a message to a channel.

Feed posts

Tool Description
list_feed_posts Recent feed posts in a workspace.
create_feed_post Create a new feed post.
comment_on_post Comment on a feed post.

Direct messages

Tool Description
list_conversations DM and group chat conversations the caller is part of.
list_conversation_messages Messages in a single conversation.
send_conversation_message Reply in a DM or group chat.

Long-form content

Tool Description
create_article Publish a new article in a workspace.
create_question Open a new Q&A question.

Support tickets

Tool Description
list_my_tickets Tickets the caller has filed.
create_ticket File a new support ticket.
comment_on_ticket Reply on a ticket thread.

Admin / moderator

These require the caller's API key to be minted by an admin/moderator (the relevant scope is checked server-side, same as the web UI).

Tool Description
list_all_tickets Tenant-wide ticket queue. Requires tickets:manage.
update_ticket Change a ticket's status / assignee. Requires tickets:manage.
list_workspace_members Members of a workspace. Requires members:view.

Security model

  • Authentication: the tenant API key acts as the caller. There's no per-tool gating in the MCP server itself — the Django API enforces scopes on every request, returning 403 when the calling key lacks the right scope. If your agent shouldn't be able to do something, mint a key with narrower scopes via the admin API-keys UI.
  • Audit: every API call lands in integrations.APIKeyUsageLog (90-day retention) just like any other key-driven call, so you can audit the agent's behaviour after the fact.

Publishing

Build + upload:

cd mcp/
pip install build twine
python -m build
python -m twine upload dist/*

Local development

cd mcp/
pip install -e ".[dev]"
pytest

Run the server manually to sanity-check tool registration:

ARYTHMATIC_API_BASE_URL=http://localhost:8000/api/v1 \
ARYTHMATIC_API_KEY=arc_test_xxxx \
python -m arythmatic_connect_mcp

(Stdio-mode — you can pipe MCP JSON-RPC at it manually if you want, but typically you wire it into an MCP client.)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

arythmatic_connect_mcp-0.1.0.tar.gz (13.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

arythmatic_connect_mcp-0.1.0-py3-none-any.whl (15.0 kB view details)

Uploaded Python 3

File details

Details for the file arythmatic_connect_mcp-0.1.0.tar.gz.

File metadata

  • Download URL: arythmatic_connect_mcp-0.1.0.tar.gz
  • Upload date:
  • Size: 13.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.2

File hashes

Hashes for arythmatic_connect_mcp-0.1.0.tar.gz
Algorithm Hash digest
SHA256 3a6c8b38420ab1a1339d8dd4964776cefa5aecd7868bc77fb684fe584ea202c1
MD5 da3781c8efc4fa65956710a36067de1b
BLAKE2b-256 00d121e245794dfb691de0a066cb7be650da3f501c7f582fd97f046c44f395de

See more details on using hashes here.

File details

Details for the file arythmatic_connect_mcp-0.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for arythmatic_connect_mcp-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 a084a6a53b8fa8e696a3c1c299fee0bd1e9d0dad28ec7253cd16529be4aecab1
MD5 646c649f70e1f39f954f227262707280
BLAKE2b-256 16ef4cdc954053fe3c6ff6f16894b2297aa0c26c3d72c7e63e45d5e4b331c4f2

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page