Skip to main content
Plyne MCP lets an AI assistant read your workspace’s Signal Agents, Sheets, Leads, Accounts, and saved signal history. Ask a question in your assistant, and it can retrieve the relevant records and the evidence behind them. The connection reads saved results. It does not run agents, perform new enrichment, change records, or send outreach. The server address is public; your workspace data requires sign-in and approval.
Server URL

1. Add Plyne to your assistant

Claude

Explore your records and evidence in a Claude conversation.

ChatGPT

Add Plyne as a custom MCP server for your ChatGPT workspace.

Cursor

Bring account and signal context into Cursor’s agent workflow.

VS Code

Use your scoped Plyne records from Copilot.

Codex

Connect through the Codex CLI and authorize in your browser.
Use the address above in your assistant’s remote MCP setup. The server uses Streamable HTTP; no local Plyne process or API key is needed for OAuth.

2. Authorize a workspace

Start the connection from your assistant. Sign in to Plyne, choose the workspace to share, and review the application’s name and requested read access. Approve the requested read access to return to the assistant, or decline to cancel the connection. Each connection belongs to the user and workspace that authorized it. Manage or remove the connection from your assistant’s connector settings.

3. Try a first question

List my Signal Agents and the Sheets they are linked to.
Then try:
Find the Accounts associated with stripe.com and show the saved findings that explain why they were discovered.
Use a company domain present in your workspace. The API Quickstart uses Stripe for illustration; it is not a shared demo dataset. Use the Tools reference for supported operations, Authentication and permissions for access details, and Examples for useful prompts.