Skip to content
Okibi logo

Okibi MCP server

Which MCP servers are Okibi's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-28

← Back to Okibi

Whose server it is
  • Registry nameai.okibi/porchverified in the MCP registry

The registry name sits under okibi.ai. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-10-28
  • Stateanswerschecked 2026-10-28
  • Tools10
  • Server calls itselfporch
  • Server version0.1.0
  • Protocol version2025-06-18

On 2026-10-28 the server answered initialize and tools/list without a sign-in. First answer in 230 ms.

Should an agent connect itNot proven either way yet
  • Whether its tools do the product's job has not been judged yet.
  • Checked on 5 of the 7 days it takes to judge reliability.

Once connected

  • create_inbox, set_response, create_approval_link, delete_inbox change data without a sign-in: anyone with the address can call them.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 5, the first on 2026-10-28
  • INITE MCP scoreafter a weeka score needs 7 daily checks: 5 so far
  • Answeredmeasuring5 of the 7 daily checks it takes to print a share
  • Median first answer241 ms
  • Space in an agent's context≈ 2,099 tokensthe definitions of all its tools (10), which an agent carries on every turn
  • Tool definitionsA · 100 of 100
  • Typed results10 of 10 toolsdeclared outputSchema: results come typed, read once
  • Every tool passes all six checks.

What an agent can do with it

  • not sure · generate a command-line tool from an existing codebase
  • not sure · let agents run product commands with structured JSON output
  • cannot · give agents a signed identity to set up accounts

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

  • Tools that change data, on a server that answers without a sign-in: create_inbox, set_response, create_approval_link, delete_inbox

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers10 tools · read 2026-10-28
  • create_inbox

    Get a public URL that records every HTTP request sent to it, plus the inbox_key that all other tools need. Use it as a webhook endpoint, OAuth redirect URI, job callback URL, or with create_approval_link. The key is shown only once: keep…

  • get_inbox

    Show an inbox: its URL, expiry, request count, latest request id, configured responses and approval links.

  • list_requests

    List captured requests as summaries, oldest first. Without "after" you get the most recent ones; with "after" you page forward from that request id. Use get_request for headers and body. Does not move wait_for_request's position.

  • get_request

    Full details of one captured request: method, path, parsed query, headers, body (UTF-8 text or base64), plus "json" or "form" when the body parses. Pass replay_to to also get a byte-exact curl command and fetch recipe for delivering it…

  • wait_for_request

    Block until a request arrives, then return it in full (same shape as get_request). By default it returns the next request not yet returned by wait_for_request for this inbox and these filters, so nothing that arrives between calls is…

  • set_response

    Choose what Porch sends back to callers, for the whole inbox or a sub-path: status, headers and body. Useful for webhook verification handshakes, or a friendly page after an OAuth redirect (or a 302 with a Location header). Pages are…

  • create_approval_link

    Make a link to send a person when you need their decision ("Deploy to production?"). It opens a page with your title, message and one button per choice. Only a button press counts, so link previews and email scanners cannot answer by…

  • wait_for_approval

    Block until the person answers an approval link, then return their choice. Returns at once if it was already answered. On timeout, call it again.

  • clear_requests

    Delete all captured requests from the inbox. The URL, key and response settings stay the same.

  • delete_inbox

    Delete the inbox and everything it captured. Its URL starts returning 404 and the key stops working.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "porch": {
      "type": "http",
      "url": "https://porch.okibi.ai/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http okibi https://porch.okibi.ai/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://porch.okibi.ai/mcp as a custom connector in the settings.

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.