Postscript MCP server for your AI

No. This is a credential paste: the connect form asks for an API key, and you copy it from the Postscript dashboard at app.postscript.io → Account → API. There is no OAuth app to register and no redirect URL to configure. The key you paste is what scopes the MCP server to a single store.

It is the optional second field, labeled Shop token (partners only, optional). If you are connecting your own store, leave it blank and the API key alone is enough. Both fields are stored as passwords.

Not directly. send_message delivers an SMS or MMS to an existing subscriber, so a new contact has to exist first — create_subscriber adds one, and get_keywords shows the opt-in words people text in to join on their own. Opt-outs stay explicit too: unsubscribe removes a subscriber from all Postscript messaging, and redact permanently erases their personal data for GDPR or CCPA.

Ask the agent for get_webhook_signing_token, which returns the signing token for the account, and check incoming payloads against it on your side. Before you go live, test_webhook sends a mock event to the callback_url subscribed to a webhook so you can confirm your endpoint handles it. If the endpoint moves or needs different headers, update_webhook changes the callback_url and custom headers by id without recreating the subscription.

It works with Flows through create_custom_event, which is the v2 replacement for v1 Triggers. An agent sends a custom event and Postscript uses it to start a Flow or branch one that is already running. That is what makes prompts like tagging a segment and then firing an event resolve in a single pass.

18 operations across 6 API modules — Webhooks, Subscribers, Keywords, Messages, Compliance and Events — split 8 read and 10 write. 14 load into your AI client by default and the remaining 4 stay searchable, so you can pull them in when a task needs them. On top of that, PopMCP can watch 7 triggers, including message received, subscriber opted in, subscriber opted out and keyword created.

No. PopMCP hosts it and gives you a Streamable HTTP MCP URL. There is nothing to deploy, no callback URL to register, and no infrastructure to keep running or patch.

Any client that supports a remote Streamable HTTP MCP server — Claude Desktop, Claude Web, ChatGPT custom connectors, Codex and the Codex CLI, OpenCode and Z.ai's ZCode, as well as editors such as Cursor, Windsurf and Zed. Your client signs in to PopMCP with OAuth; the provider credential never leaves the server.

Loading 18 tool definitions into every conversation burns context before the work starts. PopMCP loads a curated 14-tool set covering the everyday workflows and keeps the rest searchable, so your AI pulls a specialist tool only when a task needs it. You can switch a connection to the full catalog at any time.

A tool is something your AI calls — there are 18 here. A trigger is something PopMCP watches for you: 7 Postscript events that can start work on the PopMCP side. Triggers never enter your AI client's context, and none of them ask you to paste a callback URL into Postscript.