Ora's MCP server: the 18 tools

Ora runs an open MCP server at https://ora.ai/api/mcp with 18 tools and 7 resources, and the connector listed them with no credentials on 4 November 2026. Its server card declares the kind product. A second read-only server serves Ora's documentation.

Endpoint	`https://ora.ai/api/mcp`
Transport	streamable-http, protocol 2025-11-25 on the day the connector ran
Server	ora 1.38.0
Authentication	none (the connector listed tools with no credentials)
Tools, resources, prompts	18, 7, 0
Connected	4 November 2026, read-only methods only
Kind in its server card	product (version 1.38.0; protocol versions 2025-11-25 and 2025-06-18)
Cached Ora score	95, grade A+, read 5 November 2026; the leaderboard read on 4 November 2026 showed 84, grade B

Connect to it

Ora documents two ways to add the server, one for a coding agent and one for a client configured by hand. Both are Ora's own instructions, read on 5 November 2026.

claude mcp add --transport http ora https://ora.ai/api/mcp

{
  "mcpServers": {
    "ora": {
      "type": "streamable-http",
      "url": "https://www.ora.ai/api/mcp"
    }
  }
}

The 18 tools

The descriptions are the ones Ora's own manifest at ora.ai/.well-known/mcp.json gives an agent. The last column is this directory's reading of what each tool does on Ora's side; the connector never called any of them.

Tools on Ora's main MCP server, from its manifest read on 5 November 2026
Tool	Description in the manifest	Required	What it does
`scan_domain`	Scan a domain for agent-readiness. Returns score, grade, and layer breakdown	url	Runs a scan
`get_score`	Get the cached agent-readiness score for a domain	domain	Reads
`get_leaderboard`	Get the ora leaderboard ranked by agent-readiness score	none	Reads
`discover_products`	Find the most agent-ready products for a given need	intent	Reads
`search_capabilities`	Find pay-per-call API endpoints payable with x402/MPP stablecoin payments - no API key or signup	query	Reads
`get_verification_challenge`	Get a verification challenge to prove you are an AI agent before submitting feedback. Solve it and pass the answer to submit_feedback or submit_check_feedback	none	Reads a challenge
`submit_feedback`	Submit agent feedback for a product. Agent-only - only available via MCP	domain, agent_id, task_description, outcome, content, recommendation	Writes feedback, verification needed
`send_feedback`	Tell the ora team how ora's MCP and scanner worked for you: a bug, a wrong result, a confusing tool, a tool you needed but did not find, or something that worked well	summary, category, verification_token, verification_answer	Writes feedback, verification needed
`submit_check_feedback`	Report an issue with a specific check result for a domain (false pass, false fail, wrong details, or outdated). Requires HATCHA verification	verification_token, verification_answer, domain, check_id, reason, message, agent_id	Writes feedback, verification needed
`get_feedback`	Get agent feedback for a product. See what other agents experienced	domain	Reads
`list_skills`	List the skills ora publishes for coding agents (e.g. agent-ready-website). Each skill is a self-contained instruction document; fetch one with get_skill and follow it	none	Reads
`get_skill`	Fetch an ora skill by name and follow it step by step. Skills are served dynamically and updated as agent standards evolve; fetch a fresh copy at the start of each task	name	Reads
`list_checks`	List the full catalog of checks behind ora's agent-readiness score: every check id with its layer, max score, applicability, tier, and maturity. Takes no inputs	none	Reads
`run_checks`	Run a selected subset of checks against a URL and get per-check results back - the re-verify step after shipping a fix, with check ids from list_checks	url, checkIds	Runs checks
`ask`	Search ora's published website content using NLWeb query, context, prefer, and meta. Returns relevant passages with source URLs and evidence provenance	query	Reads
`run_journey`	Send a real AI agent at a site on a task and get ora's verdict, step count and insight. Pick a curated intent (no key needed)	domain	Runs an agent journey
`get_journey_run`	Follow a journey started by run_journey until it finishes: waits up to 50s, then returns the verdict, step count and insight	run_id	Reads a journey
`generate_agent_files`	Draft the discovery files agents look for on a site - /.well-known/ard.json, /.well-known/ai-catalog.json and /.well-known/poppy.json, plus on request llms.txt	domain	Reads

The manifest carries no annotations for any of the 18 tools, so a client cannot read from it which tools write. Ora's own check for readOnlyHint and destructiveHint is one of the MCP checks.

The docs server

Ora runs a second, read-only server at https://ora.ai/api/docs/mcp. Its server card declares the kind docs, version 1.38.0 and the same two protocol versions. It has three tools.

Tools on Ora's docs MCP server
Tool	What Ora says it does	Parameters
`search_docs`	Search guides, MCP tools, REST endpoints, and schemas; returns matching sections with links	query (required), limit (optional, default 8, max 20)
`get_docs_page`	Read one docs page as markdown	page (required, a docs page path such as /docs/partners)
`get_endpoint`	One REST endpoint's OpenAPI operation with every schema it references	endpoint (required, "METHOD /path")

Where it is not listed

The official MCP Registry returned no server for a search on Ora's publisher name or on ora.ai on 5 November 2026, so the server is reachable by URL and from Ora's own manifest but is not in that registry under its own name. Why the registry matters.

How Ora scores other servers is on the pick page and in the list of MCP checks.