Developers

An API key and a manifest. That's the whole integration.

Any agent that can call an API or an MCP server can use all of Padrino. Create a scoped key, point your agent at the manifest, and it knows exactly what it can do in the business.

Get an API key Supported agents

01
Create a scoped keySettings, API keys. Read only, draft only or full access, per module.
02
Share the manifestYour agent reads it and knows every operation its key allows.
03
Or add the MCP serverFor Claude, ChatGPT and other MCP clients, add Padrino as a connector with the same key.
04
Watch the recordEvery call is logged with the key that made it, and approvals are enforced server side.

# Paste into Muse, ChatGPT, Claude or Grok
You have access to my business in Padrino.
API key: mb_live_••••••••
Manifest: https://app.padrino.ai/api/ai-workforce/manifest

Read the manifest to see what you can do, then
reconcile last week's bank activity and draft
invoices for every job marked complete.

{
  "mcpServers": {
    "padrino": {
      "url": "https://mcp.padrino.ai/mcp",
      "headers": {
        "Authorization": "Bearer mb_live_••••••••"
      }
    }
  }
}

# What can this key do?
curl https://app.padrino.ai/api/ai-workforce/manifest?mine=1 \
  -H "Authorization: Bearer mb_live_••••••••"

# Draft an invoice
curl -X POST https://app.padrino.ai/api/invoices \
  -H "Authorization: Bearer mb_live_••••••••" \
  -d '{"customer":"Rivera Dental","lines":[...]}'

1,000+

operations in the manifest

modules behind one key

auth model for people and agents

100%

of agent calls on the audit log

Same rules for everyone

Agents are users, not a side door.

Agents authenticate the same way people do and pass through the same permission checks. Business rules live in Padrino, not in the prompt: entries must balance, posted records never change, and approvals are enforced by the server.

A developer working at a desk with two monitors and a notebook of diagrams

Questions

How does an AI agent connect to Padrino?

Create an API key in Padrino, choose what it may do, and give it to your agent with the manifest address. Agents that prefer MCP can add the Padrino MCP server with the same key.

What is the manifest?

A machine-readable list of every operation Padrino exposes: path, method, the scope it needs, and whether it reads or writes. With ?mine=1 it returns only what the calling key is allowed to do.

How are keys scoped?

Each key carries scopes per module and per access level, for example read accounting, draft invoices, manage the pipeline. A key can never do more than its scopes allow, and every call is logged.

Which agents work with it?

Any agent that can call an HTTP API or an MCP server: Claude, ChatGPT, Grok Bot, Muse, and agents you build yourself.

Give your AI a business to run.

Start with one connected system instead of assembling a stack of subscriptions first.

Build my Padrino Connect your AI