MCP MeasureMCP Measure

インデックス観測済み

vetted-consumer

Will a local LLM run on your hardware? GGUF quant, buy-vs-rent-vs-API cost, used-GPU prices.

使うべきか

品質と安全性

A
説明の品質
91%
スキーマの完全性
78%
命名の品質
87%
ポイズニングのリスク
100%
権限の一致
100%
プロトコルへの準拠
100%

検出事項（2）

  • LOWTool 'get_used_gpu_prices' description lacks action verbget_used_gpu_prices 内
  • LOWTool 'compare_hardware' description lacks action verbcompare_hardware 内

ツール定義とプロトコルへの準拠に関する自動分析に基づいています。

コンテキストコスト

~1,957トークン数（ツール定義）
~1.9 KB一般的なレスポンスサイズ
注意への影響は中程度（128k コンテキストの 1.53%）

これは、サーバーのツールがモデルのコンテキストに読み込まれるたびに消費されるおおよそのトークン数です。数が多いほど、ほかのタスクに使える注意が減ります。

インストール

ワンクリックインストール

これを `claude_desktop_config.json` ファイルに追加してください:

{
  "mcpServers": {
    "vetted-consumer": {
      "url": "https://vettedconsumer.com/mcp"
    }
  }
}

リモートエンドポイント

https://vettedconsumer.com/mcpstreamable-http

できること

ツール一覧

ツール（9）

🟢 読み取り専用🟡 書き込み🔴 削除⚪ 不明
⚪can_i_run_it(model, total_b, active_b, mxfp4, hardware, ...)

Will a given local LLM run on given hardware? Returns fit, the best quant that fits, theoretical tok/s, and real owner-measured tok/s where available.

入力スキーマ

{
  "type": "object",
  "properties": {
    "model": {
      "type": "string",
      "description": "Model name, e.g. 'Llama 70B', 'gpt-oss-120B', 'Qwen 32B'. Use list_models to see known names."
    },
    "total_b": {
      "type": "number",
      "description": "For an unlisted model: total parameters in billions"
    },
    "active_b": {
      "type": "number",
      "description": "For an unlisted model: active params in billions (= total for dense, less for MoE)"
    },
    "mxfp4": {
      "type": "boolean",
      "description": "True if the model ships natively in MXFP4 (e.g. gpt-oss)"
    },
    "hardware": {
      "type": "string",
      "description": "Hardware name/id, e.g. 'rtx-3090', 'Mac 128GB', 'Strix Halo'. Use list_hardware to see known ones."
    },
    "vram_gb": {
      "type": "number",
      "description": "For custom hardware: VRAM or unified memory in GB"
    },
    "bandwidth_gbps": {
      "type": "number",
      "description": "For custom hardware: memory bandwidth in GB/s"
    },
    "unified": {
      "type": "boolean",
      "description": "True for unified-memory machines (Macs, Strix Halo, CPU+RAM)"
    },
    "context": {
      "type": "number",
      "description": "Context window in tokens (default 8192)"
    },
    "kv_precision": {
      "type": "string",
      "enum": [
        "f16",
        "q8",
        "q4"
      ],
      "description": "KV cache precision (default f16)"
    }
  },
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟢recommend_quant(model, total_b, active_b, mxfp4, hardware, ...)

Which GGUF quantization to download for a model on given hardware: the full quant ladder with file size, max context, and tok/s for each, plus the recommended pick.

入力スキーマ

{
  "type": "object",
  "properties": {
    "model": {
      "type": "string",
      "description": "Model name, e.g. 'Llama 70B', 'gpt-oss-120B', 'Qwen 32B'. Use list_models to see known names."
    },
    "total_b": {
      "type": "number",
      "description": "For an unlisted model: total parameters in billions"
    },
    "active_b": {
      "type": "number",
      "description": "For an unlisted model: active params in billions (= total for dense, less for MoE)"
    },
    "mxfp4": {
      "type": "boolean",
      "description": "True if the model ships natively in MXFP4 (e.g. gpt-oss)"
    },
    "hardware": {
      "type": "string",
      "description": "Hardware name/id, e.g. 'rtx-3090', 'Mac 128GB', 'Strix Halo'. Use list_hardware to see known ones."
    },
    "vram_gb": {
      "type": "number",
      "description": "For custom hardware: VRAM or unified memory in GB"
    },
    "bandwidth_gbps": {
      "type": "number",
      "description": "For custom hardware: memory bandwidth in GB/s"
    },
    "unified": {
      "type": "boolean",
      "description": "True for unified-memory machines (Macs, Strix Halo, CPU+RAM)"
    },
    "context": {
      "type": "number",
      "description": "Context window in tokens (default 8192)"
    },
    "kv_precision": {
      "type": "string",
      "enum": [
        "f16",
        "q8",
        "q4"
      ],
      "description": "KV cache precision (default f16)"
    }
  },
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}
⚪cheapest_hardware_for_model(model, total_b, active_b, mxfp4, context)

The cheapest catalogued, buyable machine that runs a given model at Q4 with the requested context.

入力スキーマ

{
  "type": "object",
  "properties": {
    "model": {
      "type": "string",
      "description": "Model name, e.g. 'Llama 70B', 'gpt-oss-120B', 'Qwen 32B'. Use list_models to see known names."
    },
    "total_b": {
      "type": "number",
      "description": "For an unlisted model: total parameters in billions"
    },
    "active_b": {
      "type": "number",
      "description": "For an unlisted model: active params in billions (= total for dense, less for MoE)"
    },
    "mxfp4": {
      "type": "boolean",
      "description": "True if the model ships natively in MXFP4 (e.g. gpt-oss)"
    },
    "context": {
      "type": "number",
      "description": "Context window in tokens (default 8192)"
    }
  },
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟢list_models

List the local LLM model classes the tools know about (params, dense/MoE, native context).

入力スキーマ

{
  "type": "object",
  "properties": {},
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟢list_hardware

List the machines the tools know about (memory, bandwidth, price, buy link).

入力スキーマ

{
  "type": "object",
  "properties": {},
  "$schema": "http://json-schema.org/draft-07/schema#"
}
⚪cost_compare(hardware, price_usd, tdp_w, hours, tokens, ...)

Buy vs rent vs API cost to run a model locally: monthly/1y/3y totals, break-even months, and the energy cost per 1M tokens. Same math as /cost-calculator/.

入力スキーマ

{
  "type": "object",
  "properties": {
    "hardware": {
      "type": "string",
      "description": "Catalogued hardware name/id (see list_hardware), e.g. 'rtx-3090-used'"
    },
    "price_usd": {
      "type": "number",
      "description": "For custom hardware: price in USD"
    },
    "tdp_w": {
      "type": "number",
      "description": "For custom hardware: board power draw in watts"
    },
    "hours": {
      "type": "number",
      "description": "Active hours per day (default 3)"
    },
    "tokens": {
      "type": "number",
      "description": "Tokens generated per day, for the API comparison (default 300000)"
    },
    "kwh": {
      "type": "number",
      "description": "Electricity $/kWh (default 0.16)"
    },
    "rent": {
      "type": "number",
      "description": "Cloud GPU $/hour (default 0.59)"
    },
    "api": {
      "type": "number",
      "description": "API $/million tokens (default 1.0)"
    }
  },
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟢recommend_hardware(model, total_b, active_b, mxfp4, context, ...)

Ranked list of catalogued, buyable machines that run a model at the requested context, cheapest first, with an optional budget cap.

入力スキーマ

{
  "type": "object",
  "properties": {
    "model": {
      "type": "string",
      "description": "Model name, e.g. 'Llama 70B', 'gpt-oss-120B', 'Qwen 32B'. Use list_models to see known names."
    },
    "total_b": {
      "type": "number",
      "description": "For an unlisted model: total parameters in billions"
    },
    "active_b": {
      "type": "number",
      "description": "For an unlisted model: active params in billions (= total for dense, less for MoE)"
    },
    "mxfp4": {
      "type": "boolean",
      "description": "True if the model ships natively in MXFP4 (e.g. gpt-oss)"
    },
    "context": {
      "type": "number",
      "description": "Context window in tokens (default 8192)"
    },
    "kv_precision": {
      "type": "string",
      "enum": [
        "f16",
        "q8",
        "q4"
      ],
      "description": "KV cache precision (default f16)"
    },
    "budget": {
      "type": "number",
      "description": "Optional max price in USD"
    }
  },
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}
🟢get_used_gpu_prices(gpu)

Current typical used-GPU prices for local-AI rigs (eBay Browse API median asking + hand-verified, monthly).

入力スキーマ

{
  "type": "object",
  "properties": {
    "gpu": {
      "type": "string",
      "description": "Optional name/id filter, e.g. \"3090\""
    }
  },
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}
⚪compare_hardware(hardware, model, total_b, active_b, mxfp4, ...)

Side-by-side memory, bandwidth, price, and (with a model) fit + tok/s for 2 to 4 machines.

入力スキーマ

{
  "type": "object",
  "properties": {
    "hardware": {
      "type": "string",
      "description": "2 to 4 hardware names/ids, comma-separated"
    },
    "model": {
      "type": "string",
      "description": "Model name, e.g. 'Llama 70B', 'gpt-oss-120B', 'Qwen 32B'. Use list_models to see known names."
    },
    "total_b": {
      "type": "number",
      "description": "For an unlisted model: total parameters in billions"
    },
    "active_b": {
      "type": "number",
      "description": "For an unlisted model: active params in billions (= total for dense, less for MoE)"
    },
    "mxfp4": {
      "type": "boolean",
      "description": "True if the model ships natively in MXFP4 (e.g. gpt-oss)"
    },
    "context": {
      "type": "number",
      "description": "Context window in tokens (default 8192)"
    },
    "kv_precision": {
      "type": "string",
      "enum": [
        "f16",
        "q8",
        "q4"
      ],
      "description": "KV cache precision (default f16)"
    }
  },
  "required": [
    "hardware"
  ],
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}

コミュニティ

このサーバーを評価する

エビデンス

最近の観測

検証済みバージョンは記録されていませんツール 9 件
検証済みバージョンは記録されていませんツール 9 件
検証済みバージョンは記録されていませんツール 9 件