How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

90 days of the enterprise, pro and team plans free, no card needed. See the plans

OpenSmartRoute

eval-inline-recipe - Agent - OpenSmartRoute

Agentv1.0.0

eval-inline-recipe

Name: eval-inline-recipe
Author: DaveGold

Eval arm inline-recipe (Q1b — the DERIVED FIGURES recipe, shipped by one of two channels). Answers one building-data question using ONLY the eval-inline-recipe MCP server. Not for general use — spawne

by DaveGold(0) 0 installs

Free

Free account. Installing gives you the manifest plus copy-paste snippets.

See reviews

About

Claude Code subagent imported from DaveGold/mcp-metadata-demo (.claude/agents/eval-inline-recipe.md). Copyright stays with the author.

You are answering a single question from a user about Dutch buildings, using the tools available to you.

Answer the question as well as the tools allow. If the tools cannot answer it, say so plainly rather than guessing.

Return exactly this, and nothing else:

ANSWER: <your answer to the user's question, in at most four sentences> CALLS: TOOLS: <comma-separated tool names you called, in order, or "none"> PARAMS: <the arguments of each call, as compact JSON, in order, or "none">

Use it

Copy one of these into your project. Installing also returns the manifest and these snippets.

# after Install: the listing is in your workspace's routing pool - nothing else to configure
curl -s -X POST https://api.opensmartroute.ai/api/v1/route -H 'Authorization: Bearer $OSR_API_KEY' -H 'Content-Type: application/json' -d '{"text": "...", "plan": true}'
# or pin it on the OpenAI-compatible endpoint: {"model": "davegold-mcp-metadata-demo-eval-inline-recipe-subagent", ...}

Installed into a catalogue, chosen by a router

Install eval-inline-recipe and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

Photo: RDNE Stock project

For the business

What it costs to adopt - in engineering time.

Same client, same code
One base URL
The OpenAI-compatible endpoint answers your existing SDK calls; model=auto lets the router choose.
The SDKs
One URL, every vendor
No vendor lock-in
Point your OpenAI-compatible client at one address; add, swap or retire providers in the catalogue without touching application code.
Every vendor side by side
Next eligible model answers
Failover built in
When a provider is down or over quota the request moves on; the trace says it did.

Manifest

An Open Capability Manifest: the router reads it to know what this does, what it costs and when to pick it.

davegold-mcp-metadata-demo-eval-inline-recipe-subagent.ocm.jsonjson

{
  "ocm": "1",
  "id": "davegold-mcp-metadata-demo-eval-inline-recipe-subagent",
  "kind": "agent",
  "name": "eval-inline-recipe",
  "description": "Eval arm inline-recipe (Q1b — the DERIVED FIGURES recipe, shipped by one of two channels). Answers one building-data question using ONLY the eval-inline-recipe MCP server. Not for general use — spawned by the run-eval skill.",
  "publisher": "DaveGold",
  "version": "1.0.0",
  "capabilities": {
    "domains": [
      "general"
    ],
    "tags": [
      "agent-md",
      "github-claude-agents"
    ],
    "languages": [
      "en"
    ]
  },
  "quality_prior": 0.6,
  "examples": [
    "Eval arm inline-recipe (Q1b — the DERIVED FIGURES recipe, shipped by one of two channels). Answers one building-data question using ONLY the eval-inline-recipe MCP server. Not for general use — spawned by the run-eval skill."
  ],
  "primary": false,
  "metadata": {
    "source": {
      "provider": "github-claude-agents",
      "repository": "https://github.com/DaveGold/mcp-metadata-demo",
      "path": ".claude/agents/eval-inline-recipe.md",
      "ref": "36a4718a1e71c72e53f8ed18df22d39fe1939e21",
      "url": "https://github.com/DaveGold/mcp-metadata-demo/blob/36a4718a1e71c72e53f8ed18df22d39fe1939e21/.claude/agents/eval-inline-recipe.md",
      "key": "DaveGold/mcp-metadata-demo/.claude/agents/eval-inline-recipe.md"
    },
    "tools": [
      "mcp__eval-inline-recipe__get_building_profile",
      "mcp__eval-inline-recipe__get_weather_context",
      "mcp__eval-inline-recipe__render_chart",
      "mcp__eval-inline-recipe__render_table",
      "mcp__eval-inline-recipe__render_map",
      "mcp__eval-inline-recipe__get_tool_call_log"
    ]
  },
  "instructions": "You are answering a single question from a user about Dutch buildings, using the\ntools available to you.\n\nAnswer the question as well as the tools allow. If the tools cannot answer it, say\nso plainly rather than guessing.\n\nReturn exactly this, and nothing else:\n\nANSWER: <your answer to the user's question, in at most four sentences>\nCALLS: <number of tool calls you made>\nTOOLS: <comma-separated tool names you called, in order, or \"none\">\nPARAMS: <the arguments of each call, as compact JSON, in order, or \"none\">",
  "cost": {
    "context_tokens": 129
  }
}

Fetch it by URL: GET /api/v1/registry/davegold-mcp-metadata-demo-eval-inline-recipe-subagent/manifest?version=1.0.0

About

Use it

Install eval-inline-recipe and it becomes one more candidate the router can pick - when it fits.

What it costs to adopt - in engineering time.

One base URL

No vendor lock-in

Failover built in

Manifest

Reviews