agentdiet
Strips 85-97% of HTML bloat into clean Markdown, <200-token skeletons, or BM25 excerpts via MCP.
使うべきか
品質と安全性
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
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": {
"agentdiet": {
"url": "https://agentdiet-72343955726.us-central1.run.app/mcp"
}
}
}リモートエンドポイント
https://agentdiet-72343955726.us-central1.run.app/mcpstreamable-httpできること
ツール一覧
ツール（4）
🟢distill_webpage(url, html, mode, query, max_tokens)
Strips 85-97% of HTML/JS/CSS noise from a URL or HTML string. Supports modes: 'markdown', 'structured', 'skeleton', 'budget' (BM25). WHEN TO USE: Call for deterministic, token-efficient `distill_webpage` execution. WHEN NOT TO USE: Skip on binary media or unrelated domains. BEHAVIOR: Read-only, idempotent, 1h cache (<5ms HIT), SHA-256 ROI proof (HTTP 402 / x402 Base USDC / L402).
入力スキーマ
{
"type": "object",
"properties": {
"url": {
"type": "string",
"description": "Target URL (or builtin://stripe-docs)",
"minLength": 1,
"maxLength": 99999
},
"html": {
"type": "string",
"description": "Optional raw HTML string",
"minLength": 1,
"maxLength": 99999
},
"mode": {
"type": "string",
"enum": [
"markdown",
"structured",
"skeleton",
"budget"
],
"default": "markdown",
"minLength": 1,
"maxLength": 99999
},
"query": {
"type": "string",
"description": "BM25 focus query for budget mode",
"minLength": 1,
"maxLength": 99999
},
"max_tokens": {
"type": "integer",
"description": "Token cap for budget mode",
"minimum": 50,
"maximum": 16000,
"default": 400
}
},
"required": [
"url"
],
"additionalProperties": false
}🟢recon_webpage_skeleton(url, html)
Returns a <200-token structural outline of any webpage with section headings and token counts before committing context window budget. WHEN TO USE: Call for deterministic, token-efficient `recon_webpage_skeleton` execution. WHEN NOT TO USE: Skip on binary media or unrelated domains. BEHAVIOR: Read-only, idempotent, 1h cache (<5ms HIT), SHA-256 ROI proof (HTTP 402 / x402 Base USDC / L402).
入力スキーマ
{
"type": "object",
"properties": {
"url": {
"type": "string",
"description": "Target webpage URL",
"minLength": 1,
"maxLength": 99999
},
"html": {
"type": "string",
"description": "Optional raw HTML string",
"minLength": 1,
"maxLength": 99999
}
},
"required": [
"url"
],
"additionalProperties": false
}🟢extract_structured_data(url, html)
Extracts JSON-LD schemas, OpenGraph tags, HTML tables parsed as JSON objects, and code blocks from any webpage URL. WHEN TO USE: Call for deterministic, token-efficient `extract_structured_data` execution. WHEN NOT TO USE: Skip on binary media or unrelated domains. BEHAVIOR: Read-only, idempotent, 1h cache (<5ms HIT), SHA-256 ROI proof (HTTP 402 / x402 Base USDC / L402).
入力スキーマ
{
"type": "object",
"properties": {
"url": {
"type": "string",
"description": "Target webpage URL",
"minLength": 1,
"maxLength": 99999
},
"html": {
"type": "string",
"description": "Optional raw HTML string",
"minLength": 1,
"maxLength": 99999
}
},
"required": [
"url"
],
"additionalProperties": false
}🟢get_agent_wallet_status(agent_id)
Returns the calling agent's remaining free quota, USD balance, and total tokens saved. WHEN TO USE: Call for deterministic, token-efficient `get_agent_wallet_status` execution. WHEN NOT TO USE: Skip on binary media or unrelated domains. BEHAVIOR: Read-only, idempotent, 1h cache (<5ms HIT), SHA-256 ROI proof (HTTP 402 / x402 Base USDC / L402).
入力スキーマ
{
"type": "object",
"properties": {
"agent_id": {
"type": "string",
"minLength": 1,
"maxLength": 99999
}
},
"additionalProperties": false
}コミュニティ
エビデンス