Build a Dashboard for Your AI Agents with Claude Code and the ThinnestAI MCP Server

Your dashboard, not ours
The ThinnestAI console has analytics. But every business eventually wants its own view: calls next to its CRM's deals, leads by campaign, escalations by product line. With the ThinnestAI MCP server connected, your coding assistant can build that view for you.
Ask for it
Build a Next.js dashboard for our ThinnestAI workspace: today's conversations
by channel, calls with outcome and summary, and new leads. Store everything in
our Postgres and keep it up to date with webhooks.
The assistant can read the built-in dashboards guide, then write the backend, the database tables and the webhook receiver.
Or just ask for the number
You do not have to build anything to get a figure. Four read-only tools return the same numbers the console shows: usage and balance (replies this month, your allowance, your balance), agent analytics (conversations, leads, escalations, and the questions your agent could not answer), call totals (answered, missed, minutes, what calls cost) and WhatsApp totals (sent, delivered, read, failed). Ask "which questions did the support agent fail to answer this month?" and the assistant answers from your real data.
How it should be built
- Webhooks for live data. Subscribe to call results, captured leads, escalations and resolved conversations, and write each into your own tables as it arrives.
- The API to backfill. Page through conversations and calls once to fill history.
- Server side only. The API key belongs in your backend's environment, never in browser code.
- A Read-only key. A dashboard only needs to look, so give it a key that cannot change anything.
What the data looks like
Calls come with their status, timings, a summary, the details the agent collected, the full transcript and a recording link. Conversations come with their channel and messages. Leads come with name, email, phone and a note. All ids are stable, so your tables stay in sync.