Pancake

Pancake now runs an official MCP server at mcp.pancake.biz — connect once and whichever AI you already use (Claude.ai, ChatGPT/Codex, Cursor, Cowork…) can read the Pancake docs and pull real data from your page. No Node.js, no cloning a repo, and with Claude.ai you never touch a config file. This post gives you 10 copy-paste prompts that work today (weekly report, VIP customer export, which ad or post drove inbound, live numbers mid-meeting), then teaches the 5-part formula so you can write prompts for whatever you actually need.

Pancake's official MCP server bundles the entire API documentation as an offline snapshot — search endpoints, pull a single endpoint's full contract (with every $ref expanded), inspect webhook payloads, and read the 11 guides (auth, rate limits, webhook setup, best practices). Wire it once into Claude Desktop / Cursor / any MCP client and your AI can read Pancake docs on its own — no more copy-paste. Bonus: configure access tokens as env vars and the assistant can hit the live Pancake API for real (list conversations, send messages, stats), with the right token attached automatically per endpoint. This post walks from zero: install Node, clone, build, wire up Claude Desktop + Cursor, real-world prompts, read-only mode, and troubleshooting.

Full reference for the 7 Statistics endpoints in the Pancake chat API — page, user (agent KPIs), tag, engagement, ads + ads campaign, new customer. Includes a critical warning about the 3 DIFFERENT date param formats across endpoints (since/until Unix seconds, date_range string DD/MM/YYYY HH:MM:SS, date_range string DD/MM/YYYY) — the #1 pitfall when building BI dashboards or scheduled syncs.

Pancake CRM is one product within the Pancake ecosystem — with features entirely different from Pancake's multi-channel chat. Base URL crm.pancake.vn/api, authenticated via a per-workspace api_key, managing lead/account/contact/order/ticket. This article positions CRM within the Pancake suite, walks through api_key setup, the first test call, and an overview of the 6 endpoint groups + webhook.

Detailed guide to the 4 endpoint groups that manage a Pancake conversation's lifecycle — add/remove tags, assign staff, mark read/unread. Includes 2 lookup endpoints (tags + users), keyword-based auto-tag patterns, round-robin assignment, and a heads-up on Pancake's built-in round_robin_users feature.