AI integrations
Let your AI assistant call lomi. for you
Model Context Protocol (MCP) connects your IDE or AI client to the lomi. merchant API, create checkouts, list payments, debug webhooks, without hand-writing every HTTP request during integration work.
Getting started
Hosted MCP server
Use https://mcp.lomi.africa with a transport token plus your merchant API key, snippets in the dashboard under Integrations → MCP.
Cursor and Claude Desktop
Copy config from the dashboard, paste your keys, restart the client, same scopes as your API key.
Local stdio option
Run npx @lomi./mcp for local development when you prefer stdio transport over HTTP.
Not a staff admin tool
MCP is for <strong>your</strong> organization’s merchant API while building integrations; not for running lomi.’s internal ops.
What is lomi. MCP?
Payment infrastructure for francophone West Africa: hosted checkout, Mobile Money, cards, payouts, subscriptions, and developer APIs across UEMOA. MCP exposes that merchant API to AI clients via Model Context Protocol.
Key facts
- Hosted MCP server at mcp.lomi.africa with OAuth or API key authentication.
- Same scopes and permissions as your merchant REST API keys.
- Supports checkout sessions, customers, subscriptions, and webhook debugging from AI-assisted workflows.
MCP vs CLI
Use MCP when Cursor, Claude Desktop, or another MCP client is in your workflow. Use the lomi. CLI for CI pipelines and scripted automation.
How to connect MCP
- Create a lomi. merchant account and complete KYC.
- Open Dashboard → Integrations → MCP and copy the client configuration.
- Paste the config into your AI client and authenticate with your API key or OAuth.
MCP FAQ
Does MCP work in sandbox mode?
Yes. Sandbox and live keys follow the same MCP scopes as the REST API.
CLI instead?
For scripts and CI, the lomi. CLI may fit better than MCP. Use MCP when an AI assistant is in your workflow.
See pricing