for agents · for AI labs · for engineers

The version of the software the agent can use.

One install for the assistant, one signup for the business. The assistant calls typed verbs against real backends. The business sees the record the same way a human submission would land. No special agent inbox to maintain.

Install in a minute →Read the specMIT open spec
The argument, in three beats

The current web is the wrong shape for the work.

  1. 01

    Where most agents fail today

    They land on a human marketing page. They guess at buttons. They click around hoping a form is what the user asked for. They sometimes succeed; mostly they apologise and ask the user to do it themselves. The interface is the wrong shape for the worker.

  2. 02

    What the agent actually wants

    A typed contract per capability. Declared invariants. A backend that honours them. Errors that can be acted on. Idempotency where it matters. Latency that fits a sync conversation. No marketing copy in the response.

  3. 03

    What we built

    A small, shared vocabulary of verbs across CRM, scheduling, lead intake, and project work. Each verb has a spec page with input and output schemas, declared invariants, and a public score from real backend invocations every release. One MCP install reaches every business inside.

The glossary

Verbs the assistant has learned so far.

Each one is a typed contract: what the assistant gives, what it gets back, what cannot change while inside. The list extends. New verbs go through the spec, get a conformance suite, then ship. Browse the live registry at /capabilities.

  1. 01ask_availability
    scheduling
    Open slots for a service on a date. No state change.read
  2. 02book_slot
    scheduling
    Hold a slot for a named customer. Calendar blocks.write
  3. 03cancel_booking
    scheduling
    Free a held slot. Idempotent on repeat.write
  4. 04list_services
    scheduling
    What this provider offers, with prices and durations.read
  5. 05submit_request
    leads
    Land a structured inbound request. AI qualification fires automatically.write
  6. 06request_quote
    leads
    Ask for a priced quote on a defined scope.write
  7. 07start_project
    projects
    Kick off a real engagement from a brief.write
  8. 08create_ticket
    projects
    Open a support / bug / billing / change request.write
  9. 09find_contact
    crm
    Search by email or name. Returns matches with id and metadata.read
  10. 10log_activity
    crm
    Append a touchpoint to a contact timeline. Cross-tenant resolved by email.write
  11. 11pay_invoice
    crm
    Surface a payable invoice and its public Stripe checkout URL.read
Install

One config block. A minute.

The universal config runs through a local bridge that every desktop assistant understands. The direct config is for newer assistants that already speak the modern transport.

universal · works everywhere
{
  "mcpServers": {
    "whatcanido": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://whatcanido.dev/api/mcp"
      ]
    }
  }
}
direct · newer assistants
{
  "mcpServers": {
    "whatcanido": {
      "url": "https://whatcanido.dev/api/mcp"
    }
  }
}

Per-assistant guides: Claude Desktop · Claude.ai · ChatGPT · Cursor · VS Code Copilot.

Already reachable

Real tenants the assistant can walk into.

Each is a real business with services, hours, calendar, an inbox. Pick one and walk through it the way the customer would. The assistant calls find_providers get_provider_actions submit_action.

  1. June Fielding · Career Coaching
    leadkit:june-fielding

    Eight-week career resets for senior engineers thinking about the next step.

  2. North Bureau
    leadkit:north-bureau

    A small design studio for product and brand work.

  3. Ridgeline Roofing
    leadkit:ridgeline-roofing

    Honest roof work. Free estimates within 24 hours.

  4. Aurora Hair Studio
    bookio:aurora-hair-brooklyn

    Brooklyn hair studio · est. 2014

  5. Bistro Modrá Houba
    bookio:bistro-modra-houba

    Letenské sezónní bistro · café · víno

  6. Dental Vinohrady
    bookio:dental-vinohrady

    Stomatologická klinika přijímající nové pacienty

How you can trust this

You can check before you trust.

The agent web is about to fill with "trust me" claims. Ours is the opposite. Every verb is exercised against its provider's real backend, every release. Test by test. Pass or fail. Published. When a provider's score moves, the assistant sees it and routes around them. The pressure goes the right way.

Walk the scoreboard →
Questions we get

What people actually ask.

What is this for, in one sentence?+

Whatcanido is where a small business becomes reachable by AI assistants the same way it is reachable by humans on the web. One install for the assistant, one signup for the business.

Why does this need to exist?+

Each SaaS vendor that ships its own MCP server teaches the assistant a new vocabulary, with new field names. Multiply that by every vendor and you get a standard nobody wrote. Whatcanido sits one level above and gives the assistant a shared vocabulary that means what it says.

How do I add it to Claude Desktop?+

Open Claude Desktop config (Settings → Developer → Edit Config), add this to mcpServers: { "whatcanido": { "command": "npx", "args": ["-y", "mcp-remote", "https://whatcanido.dev/api/mcp"] } }. Save, restart, done.

Claude.ai or ChatGPT?+

Settings → Connectors → Add custom connector → https://whatcanido.dev/api/mcp. No OAuth required for the public surface.

What does my business actually get from being inside?+

When a customer asks an AI to do something at your business (book, ask, quote, look up an invoice), the assistant reaches you the same way a human form submission would. Your existing flows fire. You don't operate a separate agent inbox.

Is it open?+

The spec, the protocol, and the conformance test runner are MIT-licensed. The SaaS products that run on top are commercial.

How do I trust the system actually does what it says?+

Every capability is exercised against the real backend on a schedule. Every result is published on the capability detail page. If a provider stops behaving, the score moves immediately and the assistant can route around them.