Web Intelligence
Web Intelligence
Scrape, crawl, search, extract, and screenshot public web pages into agent-ready context.
Status: Preview.
Web Intelligence is the first hosted API family under Agent Mag. It is built for agents that need current public web context without building browser workers, content extraction, retries, queues, and usage ledgers from scratch.
Endpoint Family
| Endpoint | Status | Use case | Typical output |
|---|---|---|---|
POST /v1/scrape | Preview | Fetch one URL. | Markdown, text, links, metadata, optional screenshot. |
POST /v1/crawl | Planned | Traverse many pages. | Job ID, page results, crawl graph, failures. |
POST /v1/search | Planned | Find and rank URLs before capture. | Ranked results and selected page context. |
POST /v1/extract | Planned | Convert pages into typed JSON. | Schema-shaped JSON and confidence notes. |
POST /v1/screenshot | Preview Spec | Capture visual page state. | Image URL, viewport metadata, timing details. |
Shared Request Rules
All hosted endpoints use bearer auth:
Requests should include public URLs only. Private pages, credentialed browsing, and bypassing access controls are out of scope.