Koskamo

IoMarkets.ai MCP

A remote MCP server providing read-only tools for IoMarkets Federated Computing. It allows agents to retrieve overviews, guides, and term explanations, list and inspect runs with receipt verification, estimate costs, and draft evaluation requests. It does not send, buy, or start anything.

Endpoint and access

No sign-in to connect and list tools. Individual tools may still need a vendor account, credentials, or permissions; check the documented prerequisites.

IoMarkets.ai MCP logo, by IoMarkets AICommunity listingStreamable HTTPNo sign-inEndpoint checked 2026-11-01Published by IoMarkets AI
MCP endpoint (Streamable HTTP)
https://iomarkets.ai/mcp
Vendor documentation

What you can do with it

  • Verify Federated Computing receipts and access review packs
  • Estimate costs for runs on confidential hardware
  • List available runs and get detailed run information
  • Draft evaluation requests for benchmark and data owners

Source and maintenance

Community listingRegistry 1.1.0
mcp-registry: io.github.SergoVashakmadze/iomarkets-ai@1.1.0Project website

Protocol and capabilities

Server
iomarkets-ai 1.1.0
MCP protocol
2025-06-18
Tools
7 advertised
Resources
Not advertised
Resource templates
Not advertised
Prompts
Not advertised
Server-advertised capabilities
{
  "tools": {
    "listChanged": false
  }
}

Server-provided instructions

Untrusted server output shown for inspection. It is not Koskamo guidance and is excluded from knowledge search.

Read-only tools for IoMarkets.ai, a technology demonstration of verifiable AI for finance: what is built, step-by-step guides with exact button names, plain-English terms, cost estimates, and the published Federated Computing runs with their receipts checked. Nothing here sends, buys or starts anything; give the person the link and the button to press.

Tools

7 advertised

get_overview — What IoMarkets.ai is
Advertised by the server during the last successful anonymous check.
Advertised read-only

What IoMarkets.ai is and what is built: Federated Computing (private computing across organisations on confidential hardware), with the main pages and what is a demonstration.

get_overview input schema
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
get_overview annotations
{
  "readOnlyHint": true,
  "openWorldHint": false
}
get_guide — How to do something on the site
Advertised by the server during the last successful anonymous check.
Advertised read-only

Step-by-step instructions, with the exact button names, for one task on iomarkets.ai. Without a task, lists the tasks.

get_guide input schema
{
  "type": "object",
  "properties": {
    "task": {
      "enum": [
        "look-around",
        "runs",
        "verify-receipt",
        "audit-trail",
        "request",
        "review-pack",
        "pilot",
        "cost",
        "own-assistant"
      ],
      "type": "string",
      "description": "Which task. Leave out to list them all."
    }
  },
  "additionalProperties": false
}
get_guide annotations
{
  "readOnlyHint": true,
  "openWorldHint": false
}
explain_term — Explain a word
Advertised by the server during the last successful anonymous check.
Advertised read-only

A plain-English meaning of a word used on the site: digital signature, invoice, Federated Computing, confidential hardware, attestation, receipt, AI agent, MCP. Without a word, lists them all.

explain_term input schema
{
  "type": "object",
  "properties": {
    "term": {
      "type": "string",
      "description": "The word to explain."
    }
  },
  "additionalProperties": false
}
explain_term annotations
{
  "readOnlyHint": true,
  "openWorldHint": false
}
list_runs — Federated Computing runs
Advertised by the server during the last successful anonymous check.
Advertised read-only

Every Federated Computing run done on confidential hardware, newest first: its id, what it showed and the hardware it ran on.

list_runs input schema
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
list_runs annotations
{
  "readOnlyHint": true,
  "openWorldHint": false
}
get_run — One run and its receipt
Advertised by the server during the last successful anonymous check.
Advertised read-only

One Federated Computing run in detail: what it showed, the model and data, hardware, location, result, payment, and whether its receipt verifies (checked now, with the same code as the receipt checker).

get_run input schema
{
  "type": "object",
  "required": [
    "id"
  ],
  "properties": {
    "id": {
      "enum": [
        "sealed-tdx-gpu",
        "sealed-tdx",
        "gpu-llm-14b",
        "gpu-llm-7b",
        "gpu-llm",
        "join-across-banks",
        "federated-round",
        "pilot-rehearsal",
        "gpu-compute",
        "three-banks",
        "governed",
        "intel-tdx",
        "first-run"
      ],
      "type": "string",
      "description": "The run id, from list_runs."
    }
  },
  "additionalProperties": false
}
get_run annotations
{
  "readOnlyHint": true,
  "openWorldHint": false
}
estimate_cost — Estimate what runs cost
Advertised by the server during the last successful anonymous check.
Advertised read-only

What Federated Computing runs cost in confidential cloud time, at Google Cloud list prices, for the machines and run times the real runs used, plus key services. Same numbers as the cost estimator page. An estimate, not a quote.

estimate_cost input schema
{
  "type": "object",
  "required": [
    "kind"
  ],
  "properties": {
    "kind": {
      "enum": [
        "tabular-cpu",
        "tabular-tdx",
        "gpu-small",
        "gpu-llm-14b"
      ],
      "type": "string",
      "description": "tabular-cpu (credit, fraud, risk models), tabular-tdx, gpu-small, or gpu-llm-14b."
    },
    "runs": {
      "type": "number",
      "description": "How many runs. Default 1."
    },
    "months": {
      "type": "number",
      "description": "Months the key services stay up. Default 3."
    },
    "key_services": {
      "type": "number",
      "description": "Key services, one per data or model owner. Default 2."
    }
  },
  "additionalProperties": false
}
estimate_cost annotations
{
  "readOnlyHint": true,
  "openWorldHint": false
}
draft_evaluation_request — Draft an evaluation request
Advertised by the server during the last successful anonymous check.
Advertised read-only

Turn what someone wants to evaluate, in their own words, into a Federated Computing evaluation request. Returns the message and a link that opens it in their email app; nothing is sent for them, and they can also use the form. Ask for anything missing, one question at a time, before calling it.

draft_evaluation_request input schema
{
  "type": "object",
  "required": [
    "role",
    "organisation",
    "name",
    "email",
    "goal"
  ],
  "properties": {
    "goal": {
      "type": "string",
      "description": "What they want to evaluate, in a sentence or two."
    },
    "name": {
      "type": "string"
    },
    "role": {
      "enum": [
        "Benchmark or data owner",
        "Model owner",
        "Model-risk or audit team",
        "Cloud provider",
        "Research or grant partner",
        "Other"
      ],
      "type": "string",
      "description": "Their role."
    },
    "email": {
      "type": "string",
      "description": "Where the reply should go."
    },
    "assets": {
      "enum": [
        "Tabular model or data",
        "Text classifier",
        "Large language model",
        "Not sure yet"
      ],
      "type": "string",
      "description": "The kind of model or data (optional)."
    },
    "constraints": {
      "type": "string",
      "description": "Regions, confidentiality, timing (optional)."
    },
    "organisation": {
      "type": "string"
    }
  },
  "additionalProperties": false
}
draft_evaluation_request annotations
{
  "readOnlyHint": true,
  "openWorldHint": false
}

Prompts

Not advertised

No prompt definitions available.

Resources

Not advertised

No resource definitions available.

30-day availability

Aggregated from read-only remote checks; this is not a tool-execution or security guarantee.

Successful checks
1 of 1 (100%)
Average latency
359 ms
95th percentile latency
359 ms

Maintenance history

Public check and source-refresh events. Editorial and ownership records remain private.

  • · probe · passed · 359 ms · metadata changed · 7 tool(s)

Connect from your client

Each snippet registers the same endpoint; pick the client you use.

This server is operated by the named third party. Listing and endpoint checks are not an endorsement or a security audit. Review the operator's permissions, terms, and data handling before connecting. How listings are checked and corrected.

Claude Code

Run this once in a terminal. It registers the remote server. Use --scope project to write .mcp.json for the repo.
Claude Code configuration
claude mcp add --transport http iomarkets-ai-mcp https://iomarkets.ai/mcp

Codex

Run this once. Codex stores MCP servers in ~/.codex/config.toml (or .codex/config.toml in a trusted project).
Codex configuration
codex mcp add iomarkets-ai-mcp --url https://iomarkets.ai/mcp

Cursor

Merge this server into .cursor/mcp.json in your project, then enable iomarkets-ai-mcp in your MCP settings.
Cursor configuration
{
  "mcpServers": {
    "iomarkets-ai-mcp": {
      "url": "https://iomarkets.ai/mcp"
    }
  }
}

Cline

Merge this into cline_mcp_settings.json (Cline → MCP Servers → Configure).
Cline configuration
{
  "mcpServers": {
    "iomarkets-ai-mcp": {
      "type": "streamableHttp",
      "url": "https://iomarkets.ai/mcp",
      "disabled": false
    }
  }
}

VS Code

Merge this server into .vscode/mcp.json. Use MCP: List Servers to start iomarkets-ai-mcp and inspect its tools.
VS Code configuration
{
  "servers": {
    "iomarkets-ai-mcp": {
      "type": "http",
      "url": "https://iomarkets.ai/mcp"
    }
  }
}

Gemini CLI

Run this once. It writes .gemini/settings.json in the current project; pass --scope user for ~/.gemini/settings.json.
Gemini CLI configuration
gemini mcp add --transport http iomarkets-ai-mcp https://iomarkets.ai/mcp

Claude Desktop

Merge this into claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\), then restart Claude Desktop.
Claude Desktop configuration
{
  "mcpServers": {
    "iomarkets-ai-mcp": {
      "url": "https://iomarkets.ai/mcp"
    }
  }
}

Windsurf

Merge this into ~/.codeium/windsurf/mcp_config.json, then refresh MCP in Cascade. Windsurf remote servers use serverUrl, not url.
Windsurf configuration
{
  "mcpServers": {
    "iomarkets-ai-mcp": {
      "serverUrl": "https://iomarkets.ai/mcp"
    }
  }
}

Let your agent find its tools

Find servers like IoMarkets.ai MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.

Connect Koskamo Explore workflows
See all 3,922