Website Services Finder

Automatically extracts and analyzes the services a company offers from its website using AI. It supports multiple AI providers and models, making it useful for lead generation, market research, and competitive analysis.

Overview

This actor crawls provided company websites and uses a configured AI model to intelligently identify and list the services offered. You can choose from five AI providers (OpenAI, Anthropic, Groq, Gemini, Hugging Face) and over 23 specific models. An optional post-processing step can group companies by semantically similar services.

Pricing is based on a pay-as-you-go model and depends on your chosen AI provider's API costs (e.g., Groq has a generous free tier).

Key Features

• Multi-Provider AI Extraction: Choose your AI backend for flexibility and cost control.
  • Groq: Fast, cost-effective models (Llama, Mixtral, Gemma).
  • OpenAI: Industry models (GPT-4o, GPT-3.5-turbo).
  • Anthropic: Claude models (Sonnet, Haiku, Opus).
  • Google Gemini: Gemini 1.5/2.0 models.
  • Hugging Face: Open-source models (Llama, Mistral, Phi-3) via various inference providers.
• AI-Powered Analysis: Extracts and categorizes services directly from website content.
• Service Clustering (Optional): Groups companies by similar services using semantic similarity, efficiently handling large datasets.
• Flexible Configuration: Control crawling limits, use custom prompts for extraction, and configure proxies.

How to Use

Provide a list of website URLs and select an AI provider. The actor will scrape the sites, send the content to the AI for analysis, and return a structured list of services. Enable the groupServices option to cluster the results.

Input

Configure the actor using the following parameters.

Required

• urls (array[string]): The list of company website URLs to analyze.
  • Example: ["https://www.example.com", "https://www.anotherexample.com"]
• provider (string): The AI provider to use. Options: "groq", "openai", "anthropic", "gemini", "huggingface". Default is "openai".

Optional

• custom_prompt (string): A custom prompt to guide the AI's service extraction.
• model_name (string): The specific model to use. If not set, a default for the chosen provider is used.
  • Groq: e.g., llama-3.3-70b-versatile, mixtral-8x7b-32768
  • OpenAI: e.g., gpt-4o, gpt-4o-mini, gpt-3.5-turbo
  • Anthropic: e.g., claude-3-5-sonnet-latest, claude-3-5-haiku-latest
  • Gemini: e.g., gemini-1.5-pro, gemini-1.5-flash
  • Hugging Face: e.g., meta-llama/Meta-Llama-3-8B-Instruct, microsoft/Phi-3-mini-4k-instruct
• provider_name (string): Required for Hugging Face. Specify the inference provider (e.g., "Cerebras", "Together AI"). See: https://hf.co/settings/inference-providers
• maxRequestsPerCrawl (integer): Maximum pages to scrape per site. Default: 100. Use 0 for unlimited.
• groupServices (boolean): If true, performs post-scraping clustering to group companies by similar services. Results are saved to the key-value store as GROUPED_SERVICES and GROUPED_SERVICES_SUMMARY. Default: false.

Output

The primary dataset contains the extracted services for each input URL. If groupServices is enabled, the grouped results are saved to the key-value store under the keys mentioned above.

You must supply your own API keys for your chosen AI provider via the actor's environment or input secrets. Proxy configuration is also supported.