Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@RyRank rerank these docs by relevance to 'how do I reset my password'"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A Cohere-compatible reranker, pay-per-call in USDC on Base (x402), free tier, no signup.
Hosted API: https://trigeochiral.com — set it as RYRANK_API.
Tool | What |
| Rerank documents by relevance to a query. Cohere-compatible response shape ( |
Related MCP server: skim-mcp
A hybrid reranker, not a cross-encoder: strong when the candidates already share vocabulary with the query, which is the common case for RAG retrieval output and document/tool shortlists. Built for high-volume, latency- and cost-sensitive reranking — narrowing a retrieval candidate set, deduping, triage — rather than close semantic disambiguation between near-identical phrasings. Pay-per-call and about half the price is the trade for that.
Verified against the real cohere.Client, cohere.ClientV2 and voyageai.Client SDKs:
point base_url at this API and existing rerank code runs unmodified.
pip install ryrank-mcp
claude mcp add ryrank --env RYRANK_API=https://trigeochiral.com -- ryrank-mcpor with uvx (no install):
claude mcp add ryrank --env RYRANK_API=https://trigeochiral.com -- uvx ryrank-mcpOptional: RYRANK_XPAYMENT = a base64 x402 X-PAYMENT payload, to make paid calls past the
free tier. Without it you get 25 free calls/day per IP.
curl -sX POST https://trigeochiral.com/v1/rerank -H 'content-type: application/json' \
-d '{"query":"how do I reset my password","documents":["Go to Settings > Security and click Reset Password.","Our refund policy allows returns within 30 days."],"top_n":2}'Also speaks the exact Cohere v2 (POST /v2/rerank) and Voyage (POST /voyage/v1/rerank) request
and response shapes, so their official SDKs work unmodified with just a base_url override — see
https://trigeochiral.com for details.
MIT
Rerank a list of documents/passages by relevance to a query. Cohere-compatible response (results: [{index, relevance_score}]), so it drops into an existing rerank call. One pass, no GPU, no model download, NO ACCOUNT - $0.001 USDC per call over x402 (about half of Cohere Rerank) or a free tier. Documents are supplied per call and discarded - this is NOT a persistent index. Best when the candidates already share vocabulary with the query (typical RAG retrieval output); it favours shared wording over deep pairwise scoring. Up to 1000 docs per call.
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | the query to rank against | |
| top_n | No | return only the top N (default: all) | |
| documents | Yes | 2-1000 documents (strings) | |
| return_documents | No | include the document text in each result |
| Name | Required | Description |
|---|---|---|
| meta | No | |
| results | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover safety (readOnly, idempotent, non-destructive, closed-world), and the description adds substantial context beyond them: no GPU/model download/account, $0.001 USDC per call via x402 with a free tier, documents discarded per call, and the shared-wording scoring bias. This is real behavioral disclosure an agent cannot get from the schema.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loaded with the core action, then pricing/limitations in compact clauses. Dense but nearly every sentence carries operational value; the marketing-adjacent pricing comparison is slightly expendable but still decision-relevant.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
An output schema exists, so return values need not be spelled out, and the description still notes the response shape. Between the schema, annotations, and description, an agent has everything needed to invoke this correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the schema already documents all four parameters, making 3 the baseline. The description reinforces the documents limit ('Up to 1000 docs per call') but adds little else about top_n or return_documents beyond what the schema states.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb and resource: 'Rerank a list of documents/passages by relevance to a query.' It is unambiguous what the tool does, and the Cohere-compatible framing clarifies the exact operation being performed.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It gives explicit context for when to use it ('Best when the candidates already share vocabulary with the query (typical RAG retrieval output)') and an explicit exclusion ('this is NOT a persistent index'). With no sibling tools to name, this is as complete as routing guidance can get.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v1.0.0rerankScored across 1 tool
With a single tool, there is no possibility of misselection; the one operation 'rerank' is unambiguous and its scope (per-call document reordering for a query) is clearly stated in the description.
A single lowercase, snake_case-style verb name 'rerank' establishes no conflicting convention and matches the standard terminology of the reranking domain.
The server is deliberately a one-endpoint reranking service, so one tool is defensible and well-scoped rather than bloated; however, a lone tool is on the thin side compared with typical multi-tool surfaces, leaving little room for related operations.
For a rerank-as-a-service API the core operation is fully covered, including the response shape and document limits. Minor gaps exist around auxiliary operations (e.g. model/version selection, usage or quota introspection), but none block the primary workflow.
Pay-per-call browser rendering and x402 marketplace ranking for AI agents, billed via USDC.
17-model LLM gateway + 350+ data/KYB/sanctions tools. Pay-per-call USDC via x402, no API key.
Pay-per-call AI gateway: models, speech, web search, on-chain reads and NLP tools, via x402.
Pay-per-call NLP & embeddings APIs on Base via x402. No keys/signup; agents pay per call in USDC.
