Pine Assistant MCP Server
A local MCP server that lets any LLM agent manage Pine AI tasks — negotiate bills, cancel subscriptions, resolve disputes, and make phone calls on your behalf.
README
Pine Assistant MCP Server
A local MCP server that lets any LLM agent manage Pine AI tasks — negotiate bills, cancel subscriptions, resolve disputes, and make phone calls on your behalf.
Built on the pine-assistant Python SDK.
Installation
pip install pine-mcp-server
Or run directly with uvx (no install needed):
uvx pine-mcp-server
Quick Start
1. Get your Pine AI credentials
You need an access_token and user_id from Pine AI. Either:
- Sign up at 19pine.ai and retrieve your credentials, or
- Use the built-in auth tools (
pine_auth_request_code/pine_auth_verify_code) to authenticate via email.
2. Configure your MCP client
Claude Desktop — edit claude_desktop_config.json:
{
"mcpServers": {
"pine-assistant": {
"command": "uvx",
"args": ["pine-mcp-server"],
"env": {
"PINE_ACCESS_TOKEN": "your-access-token",
"PINE_USER_ID": "your-user-id"
}
}
}
}
Cursor — edit .cursor/mcp.json:
{
"mcpServers": {
"pine-assistant": {
"command": "uvx",
"args": ["pine-mcp-server"],
"env": {
"PINE_ACCESS_TOKEN": "your-access-token",
"PINE_USER_ID": "your-user-id"
}
}
}
}
If you prefer pip install, replace "command": "uvx" with "command": "pine-mcp-server" and remove the "args" field.
3. Use it
Ask your LLM agent something like:
"Use Pine AI to negotiate my Comcast internet bill. My account number is 12345."
The agent will create a session, send your request, and check back for updates.
How It Works
The server follows a "load conversation" model — like refreshing a browser page:
- Create a session —
pine_session_create - Send a message describing the task —
pine_send_message - Wait, then check what Pine replied —
pine_get_history - Start the task when ready —
pine_task_start - Check again for results —
pine_get_history
There is no real-time streaming. The agent periodically loads the conversation history to see updates, similar to refreshing the Pine web app.
Available Tools
Authentication
| Tool | Description |
|---|---|
pine_auth_request_code |
Request a verification code via email |
pine_auth_verify_code |
Verify the code and obtain credentials |
Sessions
| Tool | Description |
|---|---|
pine_session_create |
Create a new Pine session |
pine_session_list |
List sessions with optional filters |
pine_session_get |
Get details about a session |
pine_session_delete |
Delete a session |
pine_session_url |
Get the web URL to view a session |
Conversation
| Tool | Description |
|---|---|
pine_get_history |
Load conversation history (the core "refresh" tool) |
pine_send_message |
Send a text message to Pine |
pine_send_form_response |
Submit a form that Pine sent |
pine_send_auth_confirmation |
Submit an OTP/verification code |
pine_send_location_response |
Submit location coordinates |
pine_send_location_selection |
Submit a location selection |
Tasks
| Tool | Description |
|---|---|
pine_task_start |
Start task execution |
pine_task_stop |
Stop a running task |
Attachments
| Tool | Description |
|---|---|
pine_upload_attachment |
Upload a local file (bill, screenshot, etc.) |
pine_delete_attachment |
Delete an uploaded attachment |
Social & Scheduling
| Tool | Description |
|---|---|
pine_social_share |
Share results on social media to earn credits |
pine_update_call_reminder |
Update a scheduled call reminder |
Environment Variables
| Variable | Required | Description |
|---|---|---|
PINE_ACCESS_TOKEN |
Yes* | Pine AI access token |
PINE_USER_ID |
Yes* | Pine AI user ID |
PINE_BASE_URL |
No | Pine AI backend URL (default: https://www.19pine.ai) |
PINE_DEVICE_ID |
No | Stable device identifier. Recommended when the server runs as a subprocess (Claude Desktop, Cursor) — otherwise a fresh random ID is generated on each launch if ~/.pine/device_id is unwritable. |
* Not required if you authenticate at runtime using the auth tools.
Development
pip install -e ".[dev]"
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.