StructForge
Turn messy text into strict JSON schemas agents can trust (invoice, receipt, contact, resume).
使うべきか
品質と安全性
検出事項（1）
- LOWTool 'extract_table' description lacks action verbextract_table 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
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": {
"structforge": {
"url": "https://structforge.askqa.app/mcp"
}
}
}リモートエンドポイント
https://structforge.askqa.app/mcpstreamable-httpできること
ツール一覧
ツール（6）
⚪structure_text(text, schema_name)
Convert messy text to strict JSON schema. Use for emails, chat logs, unstructured input. Schemas: invoice, receipt, contact, resume.
入力スキーマ
{
"type": "object",
"properties": {
"text": {
"type": "string",
"description": "Raw text to structure"
},
"schema_name": {
"type": "string",
"default": "generic",
"description": "Target schema: invoice, receipt, contact, resume"
}
},
"required": [
"text"
],
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢parse_invoice(text)
Parse invoice text into vendor, amount, due_date, line_items (with math_check).
入力スキーマ
{
"type": "object",
"properties": {
"text": {
"type": "string",
"description": "Invoice text or PDF extract"
}
},
"required": [
"text"
],
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢extract_table(html)
Extract HTML or markdown tables into JSON rows.
入力スキーマ
{
"type": "object",
"properties": {
"html": {
"type": "string",
"description": "HTML or table content"
}
},
"required": [
"html"
],
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}⚪normalize_data(data, target_schema)
Normalize data object to a named schema (invoice, receipt, contact, resume).
入力スキーマ
{
"type": "object",
"properties": {
"data": {
"type": "object",
"additionalProperties": {},
"description": "Input data object"
},
"target_schema": {
"type": "string",
"description": "Target schema name"
}
},
"required": [
"data",
"target_schema"
],
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}⚪validate_structure(data, schema_name)
Validate JSON against a schema. Enforces field presence and types (string/number/date/array).
入力スキーマ
{
"type": "object",
"properties": {
"data": {
"type": "object",
"additionalProperties": {}
},
"schema_name": {
"type": "string"
}
},
"required": [
"data",
"schema_name"
],
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}🟢get_schema(name)
Discover schema shapes. Pass name for one schema, or omit to list all curated schemas.
入力スキーマ
{
"type": "object",
"properties": {
"name": {
"type": "string",
"description": "Schema name; omit to list all"
}
},
"additionalProperties": false,
"$schema": "http://json-schema.org/draft-07/schema#"
}推奨プロンプト
get_schemaget_schemaコミュニティ
エビデンス