AI Tooling Capability Union

The "what can my agent call right now" feed. One actor that merges the canonical MCP server registry and the OpenRouter model catalog into a single normalized dataset under a shared capability taxonomy — so an agent (or a human) can ask one question and get the servers and the models that satisfy it, cross-layer.

This is the Tier-3 play: Tier 1 indexed the MCP registry with an agentic matcher; Tier 2 indexed OpenRouter models with change detection. The Union is the novel combination — a single queryable surface over both tool layers, plus a per-capability coverage index that tells an agent exactly what tool surface is available.

What it does

1. Fetches both layers — MCP servers from registry.modelcontextprotocol.io (cursor-paginated, deduped by name) and models from openrouter.ai/api/v1/models (keyless JSON, 422+ models).
2. Normalizes each into one record shape with a shared capabilities tag set: name, id, description, capabilities, transport, remote_url, context_length, prompt_price_per_mtok, reasoning, ref_url.
3. Matches an optional plain-English taskQuery across both layers using the same intent-aware scoring — a vision model that can call a web-search tool surfaces both the model and the server.
4. Emits a capability index — per capability tag, how many MCP servers vs models support it. One row that tells an agent exactly what tool surface is available right now.

Input example

{
  "taskQuery": "a vision model that can call a web search tool",
  "capabilities": [],
  "includeMcp": true,
  "includeModels": true,
  "maxMcpServers": 300,
  "minMatchScore": 15,
  "capabilityIndex": true
}

Output example (match records, one per layer)

{
  "kind": "match",
  "layer": "mcp",
  "name": "ai.exa/exa",
  "id": "ai.exa/exa",
  "description": "Search the web with embeddings. The Exa MCP server exposes web search as a tool.",
  "capabilities": ["web", "code"],
  "transport": "streamable-http",
  "remote_url": "https://mcp.exa.ai/mcp",
  "context_length": 0,
  "prompt_price_per_mtok": 0,
  "reasoning": false,
  "match_score": 100,
  "matched_on": ["web"],
  "ref_url": "https://registry.modelcontextprotocol.io/ai.exa/exa"
}

{
  "kind": "match",
  "layer": "model",
  "name": "Nex AGI: Nex-N2-Mini",
  "id": "nex-agi/nex-n2-mini",
  "description": "A multimodal reasoning model with image understanding.",
  "capabilities": ["reasoning", "function_calling", "json_mode", "vision"],
  "transport": "api",
  "remote_url": "",
  "context_length": 262144,
  "prompt_price_per_mtok": 0.12,
  "reasoning": true,
  "match_score": 100,
  "matched_on": ["vision"],
  "ref_url": "https://openrouter.ai/models/nex-agi/nex-n2-mini"
}

Output example (capability index record):

{
  "kind": "capability_index",
  "layer": "index",
  "name": "Capability coverage",
  "description": "Union of 302 MCP servers and 422 models.",
  "coverage": { "vision": { "mcp": 5, "model": 59 }, "web_search_model": { "mcp": 0, "model": 0 } }
}

A real run returned 302 MCP servers + 422 models merged, with 536 cross-layer task matches. The capability index showed vision: mcp=5, model=59.

How to use

1. Open the Actor and paste the JSON input (or use the form fields).
2. Set taskQuery to a cross-layer task, or leave it empty to return the full union.
3. Use capabilities to hard-filter (AND), or includeMcp / includeModels to narrow a layer.
4. Click Start. Records appear in the default dataset.

Source notes

OpenRouter's public API does not expose modality flags for most models (0/422 carry an image modality), so model vision/audio caps are derived from description text — the same way MCP server tags are. The shared taxonomy normalizes image→vision so a vision filter hits both layers. MCP pagination is cursor-based (?cursor=); ?offset= is ignored.

Pricing

Pay-per-event: a small per-run start fee plus a per-record item fee. Both sources are public keyless JSON — no proxy, no per-call API cost. See the Actor's pricing tab for the current rate.

Compliance

Reads only publicly published registry/model metadata. No personal data, no login, no platform ToS conflict.

Support

Found a bug or need a field added? Open an issue on the Actor's Issues tab.