# AI Training Data Curator (`vamsi-krishna/ai-training-data-curator`) Actor Turn any public website into a clean LLM training dataset. Crawl docs, blogs, and help centers, extract readable text, filter by language, remove duplicates, and export JSON, JSONL, or CSV for fine-tuning, RAG, and AI workflows. No coding required. - **URL**: https://apify.com/vamsi-krishna/ai-training-data-curator.md - **Developed by:** [Vamsi Krishna](https://apify.com/vamsi-krishna) (community) - **Categories:** AI, Developer tools, Other - **Stats:** 2 total users, 1 monthly users, 66.7% runs succeeded, 0 bookmarks - **User rating**: No ratings yet ## Pricing from $3.00 / 1,000 saved training pages This Actor is paid per event and usage. You are charged both the fixed price for specific events and for Apify platform usage. Learn more: https://docs.apify.com/platform/actors/running/actors-in-store#pay-per-event ## What's an Apify Actor? Actors are a software tools running on the Apify platform, for all kinds of web data extraction and automation use cases. In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours, and optionally produces a well-defined JSON output, datasets with results, or files in key-value store. In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server. Actors are written with capital "A". ## How to integrate an Actor? If asked about integration, you help developers integrate Actors into their projects. You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready. The best way to integrate Actors is as follows. In JavaScript/TypeScript projects, use official [JavaScript/TypeScript client](https://docs.apify.com/api/client/js.md): ```bash npm install apify-client ``` In Python projects, use official [Python client library](https://docs.apify.com/api/client/python.md): ```bash pip install apify-client ``` In shell scripts, use [Apify CLI](https://docs.apify.com/cli/docs.md): ````bash # MacOS / Linux curl -fsSL https://apify.com/install-cli.sh | bash # Windows irm https://apify.com/install-cli.ps1 | iex ```bash In AI frameworks, you might use the [Apify MCP server](https://docs.apify.com/platform/integrations/mcp.md). If your project is in a different language, use the [REST API](https://docs.apify.com/api/v2.md). For usage examples, see the [API](#api) section below. For more details, see Apify documentation as [Markdown index](https://docs.apify.com/llms.txt) and [Markdown full-text](https://docs.apify.com/llms-full.txt). # README ## AI Training Data Curator — Turn websites into LLM training data **Build clean datasets from any public website** for LLM fine-tuning, RAG (retrieval-augmented generation), and AI model training — without writing scrapers or cleaning HTML by hand. Paste your URLs, run the Actor, and download structured records with page title, body text, language, and word count — ready for JSON, JSONL, or CSV export. --- ### What this Actor does 1. **Crawls** pages from your start URLs (and optional sitemap) 2. **Extracts** clean article-style text — navigation menus and boilerplate are filtered out 3. **Filters** by language and minimum text length 4. **Removes** near-duplicate pages automatically 5. **Exports** one row per page to your Apify dataset You get training-ready text, not raw HTML. --- ### Who is it for? - **AI & ML teams** preparing fine-tuning or pre-training corpora - **RAG builders** indexing documentation, blogs, or knowledge bases - **Researchers** collecting web text datasets at scale - **Product teams** turning help centers or marketing sites into searchable AI content --- ### Quick start (3 steps) 1. **Start URLs** — Add one or more website URLs to crawl (or paste a sitemap URL). 2. **Max pages** — Choose how many pages to collect (default: 10). 3. **Run** — Open the **Dataset** tab when finished and download as JSON, JSONL, CSV, or Excel. **Tip:** For a single site section, use **Crawl strategy → seeds-only**. To follow internal links, use **recurse** and keep **Stay within seed domains** enabled. --- ### Ready-made tasks (Apify Console) Create these public tasks under **Actor → Tasks** to improve Store discoverability: | Task | Input highlights | |------|------------------| | **Quick demo — single page** | `maxPages: 1`, `seeds-only`, `https://httpbin.org/html` | | **Documentation crawl** | `recurse`, `stayWithinDomain: true`, `language: en`, `maxPages: 100` | | **Blog archive from sitemap** | `sitemapUrl` + `maxPages: 500`, `deduplicate: true` | --- ### Store monetization (Apify Console) Under **Actor → Monetization**: 1. Turn **off** “Pay per event + platform usage” so users see predictable pricing. 2. Set per-page event price to cover platform cost (current runs are ~$0.0005/page compute + margin). 3. Enable **Store discounts** with ~10% step-down per tier (Free → Bronze → Silver → Gold). --- ### What you get (output) Each saved page becomes one dataset record: | Field | What it contains | |-------|------------------| | `url` | Page address | | `title` | Page title | | `text` | Clean extracted body text | | `language` | Detected language (e.g. `en`) | | `wordCount` | Number of words | | `author` | Author name, when detected | | `publishedDate` | Publish date, when detected | **Example record:** ```json { "url": "https://example.com/blog/getting-started", "title": "Getting Started with RAG", "text": "Retrieval-augmented generation combines search with large language models...", "language": "en", "wordCount": 842, "author": "Jane Doe", "publishedDate": "2025-03-15" } ```` *** ### Common settings | Setting | What it does | |---------|----------------| | **Start URLs** | Where crawling begins | | **Sitemap URL** | Optional — load many URLs from `sitemap.xml` | | **Max pages** | Stop after this many pages (1–100,000) | | **Language filter** | Keep only pages in one language (e.g. `en`) | | **Minimum text length** | Skip very short pages (menus, stubs) | | **Remove duplicates** | Drop near-duplicate content (recommended: on) | | **Crawl strategy** | `recurse` = follow links; `seeds-only` = only listed URLs | | **Stay within seed domains** | Do not leave the original website | | **Export rejected pages** | Optional second dataset showing filtered-out URLs | | **Proxy configuration** | Use if the site blocks automated access | *** ### Example input ```json { "startUrls": [{ "url": "https://docs.example.com" }], "maxPages": 100, "language": "en", "minTextLength": 200, "deduplicate": true, "crawlStrategy": "seeds-only", "stayWithinDomain": true } ``` *** ### Download your dataset After a run completes: 1. Go to the run in [Apify Console](https://console.apify.com) 2. Open the **Dataset** tab 3. Click **Export** → choose JSON, JSONL, CSV, or Excel Use the dataset directly in Hugging Face, OpenAI fine-tuning pipelines, vector databases, or your own ML workflow. *** ### FAQ **Can I use this for RAG?**\ Yes. The output is clean text per URL — ideal for chunking and embedding into Pinecone, Weaviate, Chroma, or similar vector stores. **Does it work on documentation sites and blogs?**\ Yes. It is designed for article-style pages: docs, blogs, news, help centers, and marketing content. **Does it remove duplicate pages?**\ Yes, by default. Near-duplicate pages are detected and only one copy is kept. **What formats can I export?**\ JSON, JSONL, CSV, and Excel from the Apify dataset. JSONL is common for LLM training pipelines. **Do I need to code?**\ No. Configure inputs in the Apify UI and download results. API access is available if you want to automate runs. **Is crawling always allowed?**\ No. You must have permission to access and use the content you crawl. See **Legal notice** below. *** ### Legal notice You are responsible for complying with each website's terms of service and applicable copyright law. Only crawl sites you are allowed to access, respect `robots.txt`, and use extracted data in line with applicable regulations. # Actor input Schema ## `startUrls` (type: `array`): Seed URLs to begin crawling. Provide at least one start URL or a sitemapUrl. ## `sitemapUrl` (type: `string`): Optional sitemap.xml URL; URLs are merged with startUrls ## `maxPages` (type: `integer`): Maximum number of pages to crawl and export. ## `maxSitemapUrls` (type: `integer`): Cap merged seed/sitemap URLs before crawling (defaults to maxPages) ## `minTextLength` (type: `integer`): Minimum extracted text length in characters; shorter pages are filtered out. ## `language` (type: `string`): ISO 639-1 code (e.g. en). Leave empty to accept all. ## `deduplicate` (type: `boolean`): When true, near-duplicate pages are removed using SimHash fingerprinting. ## `maxFingerprints` (type: `integer`): Maximum SimHash fingerprints kept in memory (FIFO eviction when exceeded) ## `crawlStrategy` (type: `string`): recurse: follow links from seed pages; seeds-only: crawl only seed/sitemap URLs ## `stayWithinDomain` (type: `boolean`): When recursing, only follow links on seed hostnames ## `includeUrlGlobs` (type: `array`): Optional glob patterns; enqueued URLs must match at least one when set ## `excludeUrlGlobs` (type: `array`): Optional glob patterns; matching URLs are never enqueued ## `maxConcurrency` (type: `integer`): Maximum number of pages fetched in parallel ## `minRequestDelaySecs` (type: `number`): Minimum delay between requests for politeness (0 = no limit) ## `maxDepth` (type: `integer`): Maximum link hops from seed URLs (0 = seeds only). Leave empty for unlimited depth. ## `exportRejectedPages` (type: `boolean`): When true, URLs filtered out during processing are written to a separate 'rejected' dataset with filterReason ## `proxyConfiguration` (type: `object`): Apify proxy settings for blocked or geo-restricted sites ## Actor input object example ```json { "startUrls": [ { "url": "https://docs.example.com/getting-started" } ], "sitemapUrl": "https://docs.example.com/sitemap.xml", "maxPages": 100, "minTextLength": 200, "language": "en", "deduplicate": true, "maxFingerprints": 50000, "crawlStrategy": "seeds-only", "stayWithinDomain": true, "maxConcurrency": 3, "minRequestDelaySecs": 0, "exportRejectedPages": false } ``` # Actor output Schema ## `dataset` (type: `string`): Crawled pages with url, title, text, language, word count, author, and published date. ## `rejected` (type: `string`): Pages filtered out during crawling (only when Export rejected pages is enabled). ## `summary` (type: `string`): Crawl statistics: pages crawled, saved, filtered, and failed. # API You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup. ## JavaScript example ```javascript import { ApifyClient } from 'apify-client'; // Initialize the ApifyClient with your Apify API token // Replace the '' with your token const client = new ApifyClient({ token: '', }); // Prepare Actor input const input = { "startUrls": [ { "url": "https://httpbin.org/html" } ], "maxPages": 1, "minTextLength": 50, "deduplicate": false, "crawlStrategy": "seeds-only", "maxConcurrency": 3 }; // Run the Actor and wait for it to finish const run = await client.actor("vamsi-krishna/ai-training-data-curator").call(input); // Fetch and print Actor results from the run's dataset (if any) console.log('Results from dataset'); console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`); const { items } = await client.dataset(run.defaultDatasetId).listItems(); items.forEach((item) => { console.dir(item); }); // 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs ``` ## Python example ```python from apify_client import ApifyClient # Initialize the ApifyClient with your Apify API token # Replace '' with your token. client = ApifyClient("") # Prepare the Actor input run_input = { "startUrls": [{ "url": "https://httpbin.org/html" }], "maxPages": 1, "minTextLength": 50, "deduplicate": False, "crawlStrategy": "seeds-only", "maxConcurrency": 3, } # Run the Actor and wait for it to finish run = client.actor("vamsi-krishna/ai-training-data-curator").call(run_input=run_input) # Fetch and print Actor results from the run's dataset (if there are any) print("💾 Check your data here: https://console.apify.com/storage/datasets/" + run["defaultDatasetId"]) for item in client.dataset(run["defaultDatasetId"]).iterate_items(): print(item) # 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start ``` ## CLI example ```bash echo '{ "startUrls": [ { "url": "https://httpbin.org/html" } ], "maxPages": 1, "minTextLength": 50, "deduplicate": false, "crawlStrategy": "seeds-only", "maxConcurrency": 3 }' | apify call vamsi-krishna/ai-training-data-curator --silent --output-dataset ``` ## MCP server setup ```json { "mcpServers": { "apify": { "command": "npx", "args": [ "mcp-remote", "https://mcp.apify.com/?tools=vamsi-krishna/ai-training-data-curator", "--header", "Authorization: Bearer " ] } } } ``` ## OpenAPI specification ```json { "openapi": "3.0.1", "info": { "title": "AI Training Data Curator", "description": "Turn any public website into a clean LLM training dataset. Crawl docs, blogs, and help centers, extract readable text, filter by language, remove duplicates, and export JSON, JSONL, or CSV for fine-tuning, RAG, and AI workflows. No coding required.", "version": "0.0", "x-build-id": "5fO8t22F8I34Cq0vV" }, "servers": [ { "url": "https://api.apify.com/v2" } ], "paths": { "/acts/vamsi-krishna~ai-training-data-curator/run-sync-get-dataset-items": { "post": { "operationId": "run-sync-get-dataset-items-vamsi-krishna-ai-training-data-curator", "x-openai-isConsequential": false, "summary": "Executes an Actor, waits for its completion, and returns Actor's dataset items in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK" } } } }, "/acts/vamsi-krishna~ai-training-data-curator/runs": { "post": { "operationId": "runs-sync-vamsi-krishna-ai-training-data-curator", "x-openai-isConsequential": false, "summary": "Executes an Actor and returns information about the initiated run in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/runsResponseSchema" } } } } } } }, "/acts/vamsi-krishna~ai-training-data-curator/run-sync": { "post": { "operationId": "run-sync-vamsi-krishna-ai-training-data-curator", "x-openai-isConsequential": false, "summary": "Executes an Actor, waits for completion, and returns the OUTPUT from Key-value store in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK" } } } } }, "components": { "schemas": { "inputSchema": { "type": "object", "properties": { "startUrls": { "title": "Start URLs", "type": "array", "description": "Seed URLs to begin crawling. Provide at least one start URL or a sitemapUrl.", "items": { "type": "object", "required": [ "url" ], "properties": { "url": { "type": "string", "title": "URL of a web page", "format": "uri" } } } }, "sitemapUrl": { "title": "Sitemap URL", "type": "string", "description": "Optional sitemap.xml URL; URLs are merged with startUrls" }, "maxPages": { "title": "Max pages", "minimum": 1, "maximum": 100000, "type": "integer", "description": "Maximum number of pages to crawl and export.", "default": 10 }, "maxSitemapUrls": { "title": "Max sitemap URLs", "minimum": 1, "maximum": 100000, "type": "integer", "description": "Cap merged seed/sitemap URLs before crawling (defaults to maxPages)" }, "minTextLength": { "title": "Minimum text length", "minimum": 0, "type": "integer", "description": "Minimum extracted text length in characters; shorter pages are filtered out.", "default": 100 }, "language": { "title": "Language filter", "type": "string", "description": "ISO 639-1 code (e.g. en). Leave empty to accept all." }, "deduplicate": { "title": "Remove duplicates", "type": "boolean", "description": "When true, near-duplicate pages are removed using SimHash fingerprinting.", "default": true }, "maxFingerprints": { "title": "Max dedup fingerprints", "minimum": 1, "type": "integer", "description": "Maximum SimHash fingerprints kept in memory (FIFO eviction when exceeded)", "default": 50000 }, "crawlStrategy": { "title": "Crawl strategy", "enum": [ "recurse", "seeds-only" ], "type": "string", "description": "recurse: follow links from seed pages; seeds-only: crawl only seed/sitemap URLs", "default": "seeds-only" }, "stayWithinDomain": { "title": "Stay within seed domains", "type": "boolean", "description": "When recursing, only follow links on seed hostnames", "default": true }, "includeUrlGlobs": { "title": "Include URL globs", "type": "array", "description": "Optional glob patterns; enqueued URLs must match at least one when set", "items": { "type": "string" } }, "excludeUrlGlobs": { "title": "Exclude URL globs", "type": "array", "description": "Optional glob patterns; matching URLs are never enqueued", "items": { "type": "string" } }, "maxConcurrency": { "title": "Max concurrency", "minimum": 1, "maximum": 200, "type": "integer", "description": "Maximum number of pages fetched in parallel", "default": 10 }, "minRequestDelaySecs": { "title": "Min request delay (seconds)", "minimum": 0, "type": "number", "description": "Minimum delay between requests for politeness (0 = no limit)", "default": 0 }, "maxDepth": { "title": "Max crawl depth", "minimum": 0, "type": "integer", "description": "Maximum link hops from seed URLs (0 = seeds only). Leave empty for unlimited depth." }, "exportRejectedPages": { "title": "Export rejected pages", "type": "boolean", "description": "When true, URLs filtered out during processing are written to a separate 'rejected' dataset with filterReason", "default": false }, "proxyConfiguration": { "title": "Proxy configuration", "type": "object", "description": "Apify proxy settings for blocked or geo-restricted sites" } } }, "runsResponseSchema": { "type": "object", "properties": { "data": { "type": "object", "properties": { "id": { "type": "string" }, "actId": { "type": "string" }, "userId": { "type": "string" }, "startedAt": { "type": "string", "format": "date-time", "example": "2025-01-08T00:00:00.000Z" }, "finishedAt": { "type": "string", "format": "date-time", "example": "2025-01-08T00:00:00.000Z" }, "status": { "type": "string", "example": "READY" }, "meta": { "type": "object", "properties": { "origin": { "type": "string", "example": "API" }, "userAgent": { "type": "string" } } }, "stats": { "type": "object", "properties": { "inputBodyLen": { "type": "integer", "example": 2000 }, "rebootCount": { "type": "integer", "example": 0 }, "restartCount": { "type": "integer", "example": 0 }, "resurrectCount": { "type": "integer", "example": 0 }, "computeUnits": { "type": "integer", "example": 0 } } }, "options": { "type": "object", "properties": { "build": { "type": "string", "example": "latest" }, "timeoutSecs": { "type": "integer", "example": 300 }, "memoryMbytes": { "type": "integer", "example": 1024 }, "diskMbytes": { "type": "integer", "example": 2048 } } }, "buildId": { "type": "string" }, "defaultKeyValueStoreId": { "type": "string" }, "defaultDatasetId": { "type": "string" }, "defaultRequestQueueId": { "type": "string" }, "buildNumber": { "type": "string", "example": "1.0.0" }, "containerUrl": { "type": "string" }, "usage": { "type": "object", "properties": { "ACTOR_COMPUTE_UNITS": { "type": "integer", "example": 0 }, "DATASET_READS": { "type": "integer", "example": 0 }, "DATASET_WRITES": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_READS": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_WRITES": { "type": "integer", "example": 1 }, "KEY_VALUE_STORE_LISTS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_READS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_WRITES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_INTERNAL_GBYTES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_EXTERNAL_GBYTES": { "type": "integer", "example": 0 }, "PROXY_RESIDENTIAL_TRANSFER_GBYTES": { "type": "integer", "example": 0 }, "PROXY_SERPS": { "type": "integer", "example": 0 } } }, "usageTotalUsd": { "type": "number", "example": 0.00005 }, "usageUsd": { "type": "object", "properties": { "ACTOR_COMPUTE_UNITS": { "type": "integer", "example": 0 }, "DATASET_READS": { "type": "integer", "example": 0 }, "DATASET_WRITES": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_READS": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_WRITES": { "type": "number", "example": 0.00005 }, "KEY_VALUE_STORE_LISTS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_READS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_WRITES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_INTERNAL_GBYTES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_EXTERNAL_GBYTES": { "type": "integer", "example": 0 }, "PROXY_RESIDENTIAL_TRANSFER_GBYTES": { "type": "integer", "example": 0 }, "PROXY_SERPS": { "type": "integer", "example": 0 } } } } } } } } } } ```