Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
- Updated setup instructions to clarify API key acquisition now includes free sign-up with email verification. - Added a new "Notes" section explaining token savings, Markdown output efficiency, and handling when usage limits are exceeded. - Clarified that if token limits are reached, raw HTML will be provided instead of Markdown, along with an instructional HTML comment.
What This Skill Does
Distil converts web pages to clean Markdown via a proxy, reducing token consumption by 60–80% compared to raw HTML. Agents get consistent access to web fetch, search, screenshots, and rendered pages without managing headers or URL construction manually.
Stripping HTML boilerplate before it hits the LLM context window cuts token costs significantly, making web-augmented agents practical at scale.
When to use it
- Fetching API documentation as Markdown for an agent to read
- Searching the web for current information during a research task
- Capturing screenshots of web pages for visual inspection
- Reading JavaScript-rendered single-page apps that require execution
- Bypassing cache to get the freshest version of a frequently updated page
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: 'What are the latest Go error handling best practices?'
- 1Constructs the search URL using DISTIL_PROXY_URL and the encoded query string
- 2Sends a GET request with X-Distil-Key and Accept: text/markdown headers
- 3Receives web search results as clean Markdown
- 4Parses the Markdown to extract relevant sources and summaries
- 5Returns a synthesized answer with source links to the user
A summary of current Go error handling best practices drawn from live web search results, with source links
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.