評価指標: 概要
上級
これはEngineering, AI分野の自動化ワークフローで、17個のノードを含みます。主にSet, Webhook, Evaluation, GoogleDrive, ExtractFromFileなどのノードを使用、AI技術を活用したスマート自動化を実現。 評価指標: 概要
前提条件
- •HTTP Webhookエンドポイント（n8nが自動生成）
- •Google Drive API認証情報
- •OpenAI API Key
- •Google Gemini API Key
使用ノード (17)
ワークフロープレビュー
ノード接続関係を可視化、ズームとパンをサポート
ワークフローをエクスポート
以下のJSON設定をn8nにインポートして、このワークフローを使用できます
{
"meta": {
"instanceId": "408f9fb9940c3cb18ffdef0e0150fe342d6e655c3a9fac21f0f644e8bedabcd9",
"templateCredsSetupCompleted": true
},
"nodes": [
{
"id": "68eca3c1-4f42-4358-8c3e-f9730a960af9",
"name": "評価中？",
"type": "n8n-nodes-base.evaluation",
"position": [
860,
100
],
"parameters": {
"operation": "checkIfEvaluating"
},
"typeVersion": 4.6
},
{
"id": "dc7ea36c-c18c-4557-ba47-866c56460a39",
"name": "OpenAI Chat Model",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
468,
290
],
"parameters": {
"model": {
"__rl": true,
"mode": "list",
"value": "gpt-4.1-mini",
"cachedResultName": "gpt-4.1-mini"
},
"options": {}
},
"credentials": {
"openAiApi": {
"id": "8gccIjcuf3gvaoEr",
"name": "OpenAi account"
}
},
"typeVersion": 1.2
},
{
"id": "fb639d23-2b51-4d5c-94a4-4356825bba52",
"name": "データセット行取得時",
"type": "n8n-nodes-base.evaluationTrigger",
"position": [
-280,
170
],
"parameters": {
"sheetName": {
"__rl": true,
"mode": "list",
"value": 82338773,
"cachedResultUrl": "https://docs.google.com/spreadsheets/d/1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y/edit#gid=82338773",
"cachedResultName": "Summarization"
},
"documentId": {
"__rl": true,
"mode": "list",
"value": "1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y",
"cachedResultUrl": "https://docs.google.com/spreadsheets/d/1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y/edit?usp=drivesdk",
"cachedResultName": "96. Evaluations Test"
}
},
"credentials": {
"googleSheetsOAuth2Api": {
"id": "XHvC7jIRR8A2TlUl",
"name": "Google Sheets account"
}
},
"typeVersion": 4.6
},
{
"id": "cf8646b6-425b-4e34-89c1-50e6fb76189d",
"name": "ユーザーへの応答",
"type": "n8n-nodes-base.noOp",
"position": [
1100,
280
],
"parameters": {},
"typeVersion": 1
},
{
"id": "83cf1608-7c98-43fa-971c-ba38017cc37f",
"name": "LLM",
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"position": [
1140,
120
],
"parameters": {
"options": {},
"modelName": "models/gemini-2.0-flash"
},
"credentials": {
"googlePalmApi": {
"id": "dSxo6ns5wn658r8N",
"name": "Google Gemini(PaLM) Api account"
}
},
"typeVersion": 1
},
{
"id": "5f688735-bbb7-4906-9c4b-f1069d095f05",
"name": "出力",
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"position": [
1260,
120
],
"parameters": {
"jsonSchemaExample": "{\n \"rating\": 1,\n \"reason\": \"Tell me the reason for being lonely\"\n}"
},
"typeVersion": 1.2
},
{
"id": "8a65a118-7b26-446f-92e2-6884386c2bcf",
"name": "出力を設定",
"type": "n8n-nodes-base.evaluation",
"position": [
1480,
60
],
"parameters": {
"outputs": {
"values": [
{
"outputName": "score",
"outputValue": "={{ $json.output.score }}"
},
{
"outputName": "score_reason",
"outputValue": "={{ $json.output.reason }}"
}
]
},
"sheetName": {
"__rl": true,
"mode": "list",
"value": 82338773,
"cachedResultUrl": "https://docs.google.com/spreadsheets/d/1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y/edit#gid=82338773",
"cachedResultName": "Summarization"
},
"documentId": {
"__rl": true,
"mode": "list",
"value": "1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y",
"cachedResultUrl": "https://docs.google.com/spreadsheets/d/1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y/edit?usp=drivesdk",
"cachedResultName": "96. Evaluations Test"
}
},
"credentials": {
"googleSheetsOAuth2Api": {
"id": "XHvC7jIRR8A2TlUl",
"name": "Google Sheets account"
}
},
"typeVersion": 4.6
},
{
"id": "883908c4-e08f-4492-abe9-a4559ad5a217",
"name": "指標を設定",
"type": "n8n-nodes-base.evaluation",
"position": [
1700,
60
],
"parameters": {
"metrics": {
"assignments": [
{
"id": "34913fdd-66e9-4581-b0bd-aa564d1a5c77",
"name": "score",
"type": "number",
"value": "={{ $json.output.rating }}"
}
]
},
"operation": "setMetrics"
},
"typeVersion": 4.6
},
{
"id": "f6e24007-39e5-47e4-88fc-a7bf3a994fa6",
"name": "トランスクリプトをダウンロード",
"type": "n8n-nodes-base.googleDrive",
"position": [
-60,
70
],
"parameters": {
"fileId": {
"__rl": true,
"mode": "url",
"value": "={{ $json.input }}"
},
"options": {},
"operation": "download"
},
"credentials": {
"googleDriveOAuth2Api": {
"id": "yOwz41gMQclOadgu",
"name": "Google Drive account"
}
},
"typeVersion": 3
},
{
"id": "767a7dc0-62e7-47b8-a9eb-84a399acdf1a",
"name": "要約エージェント",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
380,
70
],
"parameters": {
"text": "={{ $json.data }}",
"batching": {},
"messages": {
"messageValues": [
{
"message": "Summarise the top 5 highlights of this video using the provided transcript."
}
]
},
"promptType": "define"
},
"typeVersion": 1.7
},
{
"id": "c1390261-2e11-4d8c-a609-76541c3dcdf0",
"name": "ファイルから抽出",
"type": "n8n-nodes-base.extractFromFile",
"position": [
160,
70
],
"parameters": {
"options": {},
"operation": "text"
},
"typeVersion": 1
},
{
"id": "a5387932-2f2f-40b7-9811-ab856417986f",
"name": "Webhook",
"type": "n8n-nodes-base.webhook",
"position": [
-500,
-30
],
"webhookId": "088cd101-9dbc-46f2-899a-d2d91c15c1e5",
"parameters": {
"path": "088cd101-9dbc-46f2-899a-d2d91c15c1e5",
"options": {}
},
"typeVersion": 2
},
{
"id": "4671049e-5ae2-455a-b7f4-b1e9119ddfef",
"name": "Gdrive URLを取得",
"type": "n8n-nodes-base.set",
"position": [
-280,
-30
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "504e6b0b-77b9-4ec9-b2f9-3dbbb6bce953",
"name": "input",
"type": "string",
"value": "={{ $json.body.gdrive_url }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "c4a8cdb7-b462-412e-92fe-82c278375e9c",
"name": "要約を評価",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
1100,
-40
],
"parameters": {
"text": "=# Inputs and AI-generated Response\n### Transcript\n{{ $('Extract from File').item.json.data }}\n\n## AI-generated Response\n{{ $('Summarise Agent').item.json.text }}\n",
"batching": {},
"messages": {
"messageValues": [
{
"message": "=# Instruction\nYou are an expert evaluator. Your task is to evaluate the quality of the responses generated by AI models.\nWe will provide you with the user input and an AI-generated responses.\nYou should first read the user input carefully for analyzing the task, and then evaluate the quality of the responses based on the Criteria provided in the Evaluation section below.\nYou will assign the response a rating following the Rating Rubric and Evaluation Steps. Give step-by-step explanations for your rating, and only choose ratings from the Rating Rubric.\n\n# Evaluation\n## Metric Definition\nYou will be assessing summarization quality, which measures the overall ability to summarize text. Pay special attention to length constraints, such as in X words or in Y sentences. The instruction for performing a summarization task and the context to be summarized are provided in the user prompt. The response should be shorter than the text in the context. The response should not contain information that is not present in the context.\n\n## Criteria\nInstruction following: The response demonstrates a clear understanding of the summarization task instructions, satisfying all of the instruction's requirements.\nGroundedness: The response contains information included only in the context. The response does not reference any outside information.\nConciseness: The response summarizes the relevant details in the original text without a significant loss in key information without being too verbose or terse.\nFluency: The response is well-organized and easy to read.\n\n## Rating Rubric\n5: (Very good). The summary follows instructions, is grounded, is concise, and fluent.\n4: (Good). The summary follows instructions, is grounded, concise, and fluent.\n3: (Ok). The summary mostly follows instructions, is grounded, but is not very concise and is not fluent.\n2: (Bad). The summary is grounded, but does not follow the instructions.\n1: (Very bad). The summary is not grounded.\n\n## Evaluation Steps\nSTEP 1: Assess the response in aspects of instruction following, groundedness, conciseness, and verbosity according to the criteria.\nSTEP 2: Score based on the rubric."
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"typeVersion": 1.7
},
{
"id": "f994670b-3c65-434e-9bf1-35c83fd0267c",
"name": "付箋1",
"type": "n8n-nodes-base.stickyNote",
"position": [
-580,
-240
],
"parameters": {
"color": 7,
"width": 1280,
"height": 700,
"content": "## 1. Setup Your AI Workflow to Use Evaluations\n[Learn more about the Evaluations Trigger](https://docs.n8n.io/integrations/builtin/?utm_source=n8n_app&utm_medium=node_settings_modal-credential_link&utm_campaign=n8n-nodes-base.evaluationTrigger)\n\nIn this scenario, we're evaluating the LLM's ability to summarise Youtube Video Transcripts. The transcripts are stored remotely in Google Drive and the expected input is the file's URL."
},
"typeVersion": 1
},
{
"id": "87506ff3-c677-4d61-963c-26ff5bcac1e6",
"name": "付箋2",
"type": "n8n-nodes-base.stickyNote",
"position": [
720,
-240
],
"parameters": {
"color": 7,
"width": 1160,
"height": 700,
"content": "## 2. Summarization: Is the AI's summary grounded in source material?\n[Learn more about the Evaluations Node](https://docs.n8n.io/integrations/builtin/?utm_source=n8n_app&utm_medium=node_settings_modal-credential_link&utm_campaign=n8n-nodes-base.evaluation)\n\nFor this evaluation, we simply want to check if the Agent's answer was grounded the incoming transcript.\nA higher score represents a greater alignment between the transcript and the expected output, indicating that the LLM is effectively sourcing relevant and accurate content to aid the generator in producing contextually appropriate responses."
},
"typeVersion": 1
},
{
"id": "3b345083-74d2-4d4b-92a4-ab22febffb4b",
"name": "付箋3",
"type": "n8n-nodes-base.stickyNote",
"position": [
-980,
-380
],
"parameters": {
"width": 380,
"height": 840,
"content": "## Try It Out!\n### This n8n template demonstrates how to calculate the evaluation metric \"Summarization\" which in this scenario, measures the LLM's accuracy and faithfulness in producing summaries which are based on an incoming transcript.\n\nThe scoring approach is adapted from [https://cloud.google.com/vertex-ai/generative-ai/docs/models/metrics-templates#pointwise_summarization_quality](https://cloud.google.com/vertex-ai/generative-ai/docs/models/metrics-templates#pointwise_summarization_quality)\n\n### How it works\n* This evaluation works best for an AI summarization workflows.\n* For our scoring, we simple compare the generated response to the original transcript.\n* A key factor is to look out information in the response which is not mentioned in the documents.\n* A high score indicates LLM adherence and alignment whereas a low score could signal inadequate prompt or model hallucination.\n\n### Requirements\n* n8n version 1.94+\n* Check out this Google Sheet for a sample data [https://docs.google.com/spreadsheets/d/1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y/edit?usp=sharing](https://docs.google.com/spreadsheets/d/1YOnu2JJjlxd787AuYcg-wKbkjyjyZFgASYVV0jsij5Y/edit?usp=sharing)\n\n\n### Need Help?\nJoin the [Discord](https://discord.com/invite/XPKeKXeB7d) or ask in the [Forum](https://community.n8n.io/)!\n\nHappy Hacking!"
},
"typeVersion": 1
}
],
"pinData": {},
"connections": {
"83cf1608-7c98-43fa-971c-ba38017cc37f": {
"ai_languageModel": [
[
{
"node": "c4a8cdb7-b462-412e-92fe-82c278375e9c",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"5f688735-bbb7-4906-9c4b-f1069d095f05": {
"ai_outputParser": [
[
{
"node": "c4a8cdb7-b462-412e-92fe-82c278375e9c",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"a5387932-2f2f-40b7-9811-ab856417986f": {
"main": [
[
{
"node": "4671049e-5ae2-455a-b7f4-b1e9119ddfef",
"type": "main",
"index": 0
}
]
]
},
"8a65a118-7b26-446f-92e2-6884386c2bcf": {
"main": [
[
{
"node": "883908c4-e08f-4492-abe9-a4559ad5a217",
"type": "main",
"index": 0
}
]
]
},
"4671049e-5ae2-455a-b7f4-b1e9119ddfef": {
"main": [
[
{
"node": "f6e24007-39e5-47e4-88fc-a7bf3a994fa6",
"type": "main",
"index": 0
}
]
]
},
"68eca3c1-4f42-4358-8c3e-f9730a960af9": {
"main": [
[
{
"node": "c4a8cdb7-b462-412e-92fe-82c278375e9c",
"type": "main",
"index": 0
}
],
[
{
"node": "cf8646b6-425b-4e34-89c1-50e6fb76189d",
"type": "main",
"index": 0
}
]
]
},
"767a7dc0-62e7-47b8-a9eb-84a399acdf1a": {
"main": [
[
{
"node": "68eca3c1-4f42-4358-8c3e-f9730a960af9",
"type": "main",
"index": 0
}
]
]
},
"c1390261-2e11-4d8c-a609-76541c3dcdf0": {
"main": [
[
{
"node": "767a7dc0-62e7-47b8-a9eb-84a399acdf1a",
"type": "main",
"index": 0
}
]
]
},
"dc7ea36c-c18c-4557-ba47-866c56460a39": {
"ai_languageModel": [
[
{
"node": "767a7dc0-62e7-47b8-a9eb-84a399acdf1a",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"f6e24007-39e5-47e4-88fc-a7bf3a994fa6": {
"main": [
[
{
"node": "c1390261-2e11-4d8c-a609-76541c3dcdf0",
"type": "main",
"index": 0
}
]
]
},
"c4a8cdb7-b462-412e-92fe-82c278375e9c": {
"main": [
[
{
"node": "8a65a118-7b26-446f-92e2-6884386c2bcf",
"type": "main",
"index": 0
}
]
]
},
"fb639d23-2b51-4d5c-94a4-4356825bba52": {
"main": [
[
{
"node": "f6e24007-39e5-47e4-88fc-a7bf3a994fa6",
"type": "main",
"index": 0
}
]
]
}
}
}よくある質問
このワークフローの使い方は？
上記のJSON設定コードをコピーし、n8nインスタンスで新しいワークフローを作成して「JSONからインポート」を選択、設定を貼り付けて認証情報を必要に応じて変更してください。
このワークフローはどんな場面に適していますか？
上級 - エンジニアリング, 人工知能
有料ですか？
このワークフローは完全無料です。ただし、ワークフローで使用するサードパーティサービス（OpenAI APIなど）は別途料金が発生する場合があります。
関連ワークフロー
AIエージェントの応答正確性をOpenAIとRAGASメソッドで評価
OpenAIとRAGASメソッドを使用してAIエージェントの応答の正確性を評価する
Set
Code
Merge
+
Set
Code
Merge
27 ノードJimleuk
エンジニアリング
AIエージェントの応答関連性をOpenAIとコサイン類似度で評価
OpenAIとコサイン類似度を使用してAIエージェントの応答の関連性を評価する
Set
Code
Evaluation
+
Set
Code
Evaluation
20 ノードJimleuk
エンジニアリング
OpenAIを使ったRAG応答の精度評価：ドキュメント・ベースライン指標
OpenAIによるRAG応答の精度評価：ドキュメントベースライン指標
Set
Evaluation
Http Request
+
Set
Evaluation
Http Request
25 ノードJimleuk
エンジニアリング
評価指標: 応答の類似性
評価指標: 応答の類似性
Set
Code
Merge
+
Set
Code
Merge
21 ノードJimleuk
エンジニアリング
基于动态提示与AirtableのAIデータ抽出
経由动态提示与Airtable实现AIデータ抽出
Set
Code
Filter
+
Set
Code
Filter
51 ノードJimleuk
人工知能
動のプロンプトとBaserowを活用したAIデータ抽出
動のプロンプトとBaserowを活用したAIデータ抽出
Set
Code
Filter
+
Set
Code
Filter
45 ノードJimleuk
人工知能
ワークフロー情報
難易度
上級
ノード数17
カテゴリー2
ノードタイプ12
作成者
Jimleuk
@jimleukFreelance consultant based in the UK specialising in AI-powered automations. I work with select clients tackling their most challenging projects. For business enquiries, send me an email at hello@jimle.uk LinkedIn: https://www.linkedin.com/in/jimleuk/ X/Twitter: https://x.com/jimle_uk
外部リンク
n8n.ioで表示 →
このワークフローを共有