Web Fetch API

The Web Fetch API extracts content from a target URL and returns the result in a format that is easier for agents, RAG systems, and content workflows to process. It can return HTML, plain text, or Markdown, and it supports selectors, readability options, cache control, custom user agents, and request cookies.

Endpoint

POST https://openapi.felo.ai/v2/web/extract

Authentication

Send your API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

See the Authentication Guide for details on creating and managing API keys.

Use with AI Agents

Install the complete felo-ai package when you want an Agent to extract webpages alongside every other available Felo Skill. The installer dynamically discovers every package directory containing SKILL.md, so it never relies on a fixed list. Install all discovered Skills in Codex ~/.codex/skills, Claude Code ~/.claude/skills, OpenClaw ~/.agents/skills, or Hermes Agent ~/.hermes/skills. Use the manual AI Agent setup guide for each platform's safe installation and API key steps.

Request

Headers

Body Parameters

Response

Success Response

The response uses a transparent pass-through design. The exact fields inside data can vary based on output_format and other extraction options.

{
  "status": "ok",
  "message": null,
  "data": {
    "content": "# Example Article\n\nExtracted article content..."
  }
}

Response Fields

Error Responses

Examples

Basic Extraction

curl -X POST https://openapi.felo.ai/v2/web/extract \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com"
  }'

Extract Markdown With Readability

curl -X POST https://openapi.felo.ai/v2/web/extract \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/article",
    "output_format": "markdown",
    "with_readability": true
  }'

Extract a Specific CSS Selector

curl -X POST https://openapi.felo.ai/v2/web/extract \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/article",
    "target_selector": "article.main-content",
    "output_format": "text"
  }'

Wait for Dynamic Content

curl -X POST https://openapi.felo.ai/v2/web/extract \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/app",
    "crawl_mode": "fine",
    "wait_for_selector": "#loaded-content",
    "output_format": "markdown",
    "timeout": 30000
  }'

Use Cases

• Give an agent webpage context before answering a question.
• Extract article text for RAG ingestion or knowledge-base creation.
• Convert webpages into Markdown for archiving or documentation workflows.
• Pull focused content from a page by using CSS selectors.
• Reduce page noise with readability processing before downstream AI analysis.

Notes

• The API handles one URL per request.
• data is passed through from the downstream extraction service, so additional fields may appear over time.
• fine mode may be slower than fast, but can be useful when the page needs richer rendering behavior.
• Some pages may block extraction, require cookies, or need a custom user_agent.