glimglim.shdocs
Tools

Web

Semantic search across the open web, and full-page extraction to clean markdown — with server-side SSR/SPA rendering, PDF support, paywall detection and residential-proxy escalation when a site fights back.

Tools & pricing

  • glim_web_search / POST /api/v1/web/search — $0.01 per call. Semantic search (Exa): query, optional include_domains / exclude_domains / published_within_days.
  • glim_web_fetch / POST /api/v1/web/fetch — $0.002 per call. url plus optional format (markdown | html | json) and selector to scope the result.

Example

request
curl -X POST https://glim.sh/api/v1/web/fetch \
  -H "x-api-key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://modelcontextprotocol.io/docs/getting-started/intro"}'
response (real call, content trimmed)
{
  "status": 200,
  "content": [
    "Get started\n\n# What is the Model Context Protocol (MCP)?\n\nMCP (Model Context Protocol) is an open-source standard for connecting AI applications to external systems. …"
  ],
  "url": "https://modelcontextprotocol.io/docs/getting-started/intro",
  "outcome": "success",
  "title": "What is the Model Context Protocol (MCP)? - Model Context Protocol",
  "total_chars": 2493
}

Good to know

  • Read outcome to classify a fetch: success, teaser, paywall, bot_challenge, login_wall, not_found and more — your agent can branch on it.
  • Fetch auto-detects SSR data (__NEXT_DATA__ etc.) and SPA shells, escalating to browser rendering only when needed — you always pay the same flat price.
  • Very large pages are truncated inline with a download_full_url to the complete extraction.

Prompts to try

  • "Search the web for recent embedding-model benchmarks, fetch the two best write-ups, and extract the numbers."
  • "Fetch this article and give me the key claims with quotes."