Give Hermes Agent eyes on the web
WebCrawler API turns any URL into clean markdown your Hermes skills can act on — instantly. Scrape a page or crawl an entire site. Proxies, JS rendering, and CAPTCHAs handled for you.

$ hermes skills install https://raw.githubusercontent.com/WebCrawlerAPI/skills/refs/heads/master/webcrawlerapi/SKILL.md --forceBuilt for AI agents
Everything your Hermes skill needs to read the web
One API. Any URL. Clean data. Ship your web research skill in minutes.
Markdown extraction
Clean markdown, ready for your skill
We load the page, strip menus, cookie banners, ads, and navigation, then return structured markdown. Feed it directly into Hermes's memory or use it in a skill response — no extra cleanup step.
- Noise removed for you. Menus, footers, ads, cookie notices stripped before the response leaves our servers.
- Works with any site. JS-rendered SPAs, paywalled previews, or plain HTML — we handle rendering with headless Chromium.
Infrastructure
No scraping infra to manage
Proxies, retries, headless browsers, CAPTCHAs, and anti-bot protection are all handled by our stack. Your Hermes skill stays simple — one API call, clean response.
- Global proxy network. Residential and datacenter proxies rotated automatically per request.
- Smart routing. Each request sent through the fastest path that can actually get the page.
Frequently Asked Questions
Common questions from Hermes Agent skill builders
How do I give my Hermes Agent access to a website?
Install the skill with hermes skills install https://raw.githubusercontent.com/WebCrawlerAPI/skills/refs/heads/master/webcrawlerapi/SKILL.md --force, then get your API key from the dashboard. After that your agent can scrape any URL and receive clean markdown — no DOM parsing, no browser setup required.
Can Hermes crawl an entire docs site, not just one page?
Yes. Use the /crawl endpoint with a starting URL and optional depth limit. We discover and fetch all linked pages, returning clean markdown for each — ideal for indexing documentation into Hermes's long-term memory.
Does it handle JavaScript-heavy sites?
Yes. We use headless Chromium to render pages before extraction, so JS-rendered content, SPAs, and dynamic dashboards all work. No browser instance needed in your skill.
Do I need a subscription?
No subscription required. Pay-as-you-go starts at $0.002 per page — you only pay for successful requests. Subscription plans are available for higher volumes at lower per-page rates.
Can I try it before installing the skill?
Yes — sign up at dash.webcrawlerapi.com, no credit card required.