OpenRouter AI Models Directory Scraper — Extract LLM Pricing & Capabilities

The OpenRouter AI Models Directory Scraper extracts structured data for every AI model available on the OpenRouter platform. Whether you are building a model comparison tool, monitoring price changes across providers, or cataloging LLM capabilities for your organization, this Actor delivers clean, structured data with zero configuration.

OpenRouter aggregates hundreds of large language models and multimodal AI providers behind a single API. Keeping track of which models support vision, audio, or tool calling — and what each costs per token — is a manual, time-consuming process. This Actor automates that entirely.

Use Cases

• LLM Comparison Dashboards — Build internal or public tools that let users filter models by context length, modality, or price.
• Price Monitoring — Track how per-token pricing evolves across providers and alert stakeholders when cheaper alternatives appear.
• Model Catalogs — Populate your product with up-to-date model metadata including supported parameters, tokenizers, and moderation flags.
• Capacity Planning — Understand max completion tokens and context windows to match the right model to your workload.

Input

Field	Type	Required	Description
maxItems	Number	No	Maximum models to return (default: 100, max: 500)
includePricing	Boolean	No	Include per-token pricing data (default: true)
includeArchitecture	Boolean	No	Include modality, tokenizer, and I/O types (default: true)
proxyConfiguration	Object	No	Proxy configuration for requests

Output

The Actor outputs a dataset with the following fields for each model:

{
  "id": "anthropic/claude-3.5-sonnet",
  "name": "Anthropic: Claude 3.5 Sonnet",
  "description": "Claude 3.5 Sonnet raises the industry bar...",
  "context_length": 200000,
  "modality": "text+image->text",
  "input_modalities": ["text", "image"],
  "output_modalities": ["text"],
  "tokenizer": "Claude",
  "instruct_type": null,
  "prompt_price": "0.000003",
  "completion_price": "0.000015",
  "image_price": "0.000004",
  "audio_price": null,
  "web_search_price": null,
  "max_completion_tokens": 8192,
  "is_moderated": false,
  "supported_parameters": ["max_tokens", "temperature", "top_p", "tools"],
  "knowledge_cutoff": "2024-04",
  "expiration_date": null,
  "created": 1718304000,
  "hugging_face_id": null
}

Pricing

Pay per event: $0.001 per model extracted. A typical run extracting 100 models costs $0.10.

Limitations

• Data reflects the current state of the OpenRouter public API. Deprecated models may still appear until they are removed by OpenRouter.
• Pricing fields are returned as strings and represent per-token costs in USD.
• The Actor makes a single API call to OpenRouter. Rate limits are respected through built-in retry logic.

FAQ

Q: Do I need an OpenRouter API key? A: No. The models directory endpoint is public and requires no authentication.

Q: How often is the data updated? A: The Actor fetches live data on every run. Run it on a schedule to keep your catalog current.

Q: Can I filter models by provider? A: The Actor returns all models. Filter the resulting dataset by the id prefix (e.g., anthropic/, openai/) in your downstream pipeline.

Changelog

• v1.0.0 — Initial release. Extracts model metadata, pricing, and architecture from the OpenRouter public API.