Atlas Agent

How to build an MCP client

Whether you ship a Custom GPT, Slack bot, or CRM widget, the client responsibilities stay the same.

Clients orchestrate user intent, tool discovery, and consent prompts. This guide illustrates the flow using a TypeScript client that powers a web intake form and Slack assistant.

Discover the catalog

Clients start by calling GET /catalog. Cache the response and refresh when the server sends an update event.

ts
const catalog = await fetch(`${SERVER_URL}/catalog`, {
  headers: { Authorization: `Bearer ${token}` },
}).then((res) => res.json());
Fetch the tool catalog

Invoke tools with tracker metadata

Send actor_id, role, workspace_id, and transaction references in each call. Capture returned event_ids so you can show status in the client UI.

Surface artifacts and alerts

Display artifact URIs inline and link to the Knowledge Base for provenance. Bubble tracker alerts when users approach budgets.

Atlas Agent

Run your own MCP server

Use Atlas Server Builder to launch faster, then upgrade to Hosted Atlas Bot when you need more seats and surfaces.

Related guides

What is an MCP server?

Pillar primer on Atlas Agent MCP strategy.

Go to What is an MCP server?

Knowledge Base

Identity, consent, provenance, tracker guard details.

Go to Knowledge Base

Pricing

Compare Server Builder, Starter Kits, and Hosted Atlas Bot.

Go to Pricing
Atlas Agent