Search API for AI

One HTTP request returns structured search results — real-time web retrieval for AI apps, RAG and agents. An alternative to Firecrawl · Exa · Tavily.

Quick start

GET /api/v1/search

curl "https://tutusoo.com/api/v1/search?q=rust&limit=5"

Query parameters

Parameter	Type	Description
q*	string	Search query (required); multiple words default to AND (all must match)
limit	int	Result count, default 10, max 100
offset	int	Pagination offset, default 0, max 990
lang	string	Language filter, e.g. zh / en
time_range	string	Indexed-time range: 1d/7d/30d/1y/all (by crawled_at)
site	string	Restrict to a site domain
lang_priority	string	Language priority (soft boost, not a hard filter): e.g. zh nudges matching-language results up without hiding others; independent of lang

Query operators (write inside q)

site:domain	Restrict to a site, e.g. site:github.com rust
intitle:word	Match only titles containing the word
"exact phrase"	Every word in the phrase must appear
-exclude	Exclude results containing the word, e.g. rust -tutorial
A OR B	Match either term, e.g. rust OR golang
filetype:type	Filter by file type (pdf/doc/ppt/xls), e.g. filetype:pdf

Response

{
  "total": 586,
  "took_ms": 142,
  "results": [
    {
      "id": "a1b2c3d4-…",         // page id for local results; may be empty for external fallback
      "url": "https://rust-lang.org/",
      "title": "Rust Programming Language",
      "snippet": "A language empowering everyone to build reliable software…",
      "domain": "rust-lang.org",
      "site_name": "Rust Programming Language", // display name (optional; client falls back to domain)
      "lang": "en",
      "score": 12.4,
      "source": "local",          // local (our index) | brave (web fallback)
      "has_llms_txt": true,
      "published_at": "2010-01-25T…", // article publish time (nullable; returned when known; RAG freshness/date)
      "crawled_at": "2026-07-18T…" // crawl time (freshness; time_range filters on this)
    }
  ]
}

site_name: display name; may be omitted — falls back to domain. source: local = our index / brave = web fallback when local results are insufficient. has_llms_txt: whether the target site ships an llms.txt (AI-friendly).

Fetch full text (RAG)

GET /api/v1/content/{id}

When a search result has source=local and a non-empty id, fetch its full body text to feed an LLM. External fallback results have no local body and cannot use this endpoint.

curl "https://tutusoo.com/api/v1/content/<local_page_id>"

# Response (RAG-relevant fields only, no internal columns):
{
  "id": "…",
  "url": "https://rust-lang.org/",
  "title": "Rust Programming Language",
  "body_text": "…full article text…",
  "author": "Jane Doe",        // author (nullable; RAG citation/byline)
  "published_at": "2026-01-…", // original publish time (nullable; freshness)
  "lang": "en",
  "word_count": 1200,
  "crawled_at": "2026-07-18T…",
  "has_llms_txt": true
}

Access & rate limits

The public search and content endpoints need no login or API key — try them directly; the public tier applies fair per-client-IP rate limiting and returns 429 when exceeded. Dedicated API keys are provisioned by the operator for confirmed integrations; this site does not show a self-serve signup that doesn't exist.

Dedicated-key responses return X-RateLimit-Limit, X-RateLimit-Remaining and X-RateLimit-Reset. On a 429, back off per the response hints — do not retry immediately without bound.