InfiniteRanks resources

Read-only MCP access to InfiniteRanks evidence

Give an assistant access to saved AI search evidence without granting it permission to change a workspace or spend collection credits.

InfiniteRanks editorial team

Four bounded tools

The MCP endpoint exposes workspace_get, prompts_list, answers_list and sources_list. Workspace context includes brand metadata. Prompt records include the latest saved revision and tags. Answers retain capture provenance and citation status. Sources retain the answer IDs that make their origin traceable. These tools read persisted records; they do not ask an AI provider a new question.

Configure the connection

Use POST https://app.infiniteranks.com/api/mcp with a manually supplied Authorization: Bearer key. The protocol version is 2025-11-25. After initialization, send MCP-Protocol-Version: 2025-11-25. Accept must include application/json and text/event-stream, although this implementation returns stateless JSON responses. OAuth, server sessions, event subscriptions and write tools are outside this compatibility subset.

Paginate and inspect provenance

List tools accept limit from 1 to 100, defaulting to 50, and an optional cursor. Continue with nextCursor until it is null. Treat cursors as opaque and keep them with the same workspace and resource. An assistant should cite the stored prompt, answer time and source URL when explaining an observation. It should not infer a visibility score from an incomplete first page.

Keep keys scoped and revocable

Create a read-only key for the intended workspace and client. Copy it only into trusted server or client configuration. Do not place it in a public page, browser bundle, repository or prompt. Revoke it if it is exposed or no longer needed. The request allowance is shared at workspace scope, so a client should handle rate-limit responses rather than polling continuously.

Sources and further reading

Sources explain the underlying standards. They do not endorse InfiniteRanks or guarantee a ranking outcome.

Be part of the conversation.

Explore a model response with your own provider key.