Skip to main content
Phantom is MealOps’ ordering infrastructure for AI agents: a single, safe interface to real restaurants — discovery, live menus, price quotes, payments, order submission, and tracking — with each restaurant’s own rules (hours, lead times, order caps) enforced server-side.

​
Pick your path

Agent Onboarding

Connect your agent: credentials and scopes, the MCP endpoint, the canonical order flow, the ordering-time model, payments, and rate limits.

MCP Tool Reference

The complete contract for all 17 Phantom MCP tools — inputs, outputs, envelopes, and the behavior rules clients must follow.

Talk to us

Partnership, credentials, scope changes, rate-limit increases, and onboarding help: support@mealops.ai

​
Endpoints at a glance

Authenticate with Authorization: Bearer <your-agent-api-key>. Access is by partnership — email support@mealops.ai to get set up.

​
Connect your agent to these docs

These docs are readable by your coding agent or assistant, so it can answer Phantom questions against the current contract instead of guessing. Note that this is a documentation server — searching these pages, not placing orders. Ordering happens over the Phantom MCP endpoint above, with your agent API key. Point a client at the docs MCP server:
No key needed — it’s public, and it exposes a search tool over every page here. In Claude Code:
Or grab a single page. The menu at the top of every page copies it as Markdown, opens it in Claude or ChatGPT with the page as context, or installs this server in Cursor or VS Code. Appending .md to any page URL does the same thing by hand — https://docs.mealops.ai/guides/agent-onboarding.md. Or feed it the whole site. /llms.txt indexes every page; /llms-full.txt is the full text of all of them in one file, for pasting into a context window or a retrieval pipeline.