Grounding with Google Search alternative

Your search tool in ten lines

Define one tool that calls SpecStream and returns the results as text. Any model that calls tools can use it.

import requests

def web_search(q: str) -> str:
    """The search tool your model calls. One call is one search at $0.005."""
    reply = requests.get(
        "https://api.specstream.ai/customsearch/v1",
        params={"key": SPECSTREAM_KEY, "q": q, "num": 10},
    ).json()
    return "\n\n".join(
        f"{i['title']}\n{i['link']}\n{i['snippet']}" for i in reply.get("items", [])
    )

# Give the text back to Gemini, Claude, GPT or any other model as the tool result.
print(web_search("Gemini 3 release date"))

The same tool in JavaScript:

async function webSearch(q) {
  const params = new URLSearchParams({ key: SPECSTREAM_KEY, q, num: "10" });
  const reply = await fetch(`https://api.specstream.ai/customsearch/v1?${params}`).then((r) => r.json());
  return (reply.items ?? []).map((i) => `${i.title}\n${i.link}\n${i.snippet}`).join("\n\n");
}

You see every result the model sees

Grounding hands you Gemini's finished answer and search suggestions HTML. SpecStream hands you the results themselves, so you choose what goes into the prompt, log what the model read, and show your users the links.

$5 per 1,000, billed for the searches you send

Grounding bills each search query the model decides to run, at $14 per 1,000 after 5,000 free a month, and stops at 1,500 requests a day per project by default. On SpecStream your code makes the call, a search of up to 10 results costs $0.005, and one account can send up to 8,800 a minute. A call stopped by a 429 charges nothing.

Page text for the prompt in the same call

On the same key, GET https://api.specstream.ai/v1/web-search?q=...&passages=true returns the text of the top pages closest to the question, each with its url and title, and an answer with the page it was read from. Your model reads the passage without a second call, and the call is priced as any search.

The most accurate web search API in the world

On SimpleQA Verified the SpecStream Search API scores 97.9 F-score on all 1,000 official questions, 978 of them correct. On WideSearch it scores 64.0 item F1 and 43.4 row F1 on all 200 official tasks. The Extract API scores 0.908 F1 on WCXB. Every score and its method is on the benchmarks page.

About one second a search, measured under load

The Custom Search route runs on the fast search. On November 9, 2026 production took 1, 5, 10 and 20 searches a second on the fast search and answered 718 of 720. The median was 0.82 s at 1 search a second and 1.07 s at 20 a second, and the 95th percentile stayed between 1.09 and 1.73 s. One account can send up to 8,800 searches a minute; a call over that gets a 429 with Retry-After and is not charged.

You decide what is kept, on each key

Create a key with zero_retention and a search made with it keeps no request text and no answer after the reply. It is not in the history and its log lines carry neither. The charge keeps the account, the key, the request ID, the price and the time. spend_cap_cents is the most the key can spend.

curl -X POST https://api.specstream.ai/v1/account/keys \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
  -d '{"name": "production", "scopes": ["account:read", "finds:spend"], "spend_cap_cents": 500000, "zero_retention": true}'

Search, pages and rows on one key

The same key reads up to 20 pages or documents a call with the Extract API at $0.80 per 1,000 pages, and a page with no text is not charged. Deep Search takes a request in plain words and returns every match on the live web as structured rows, at 5¢, 15¢ or $1 a row, plus 5¢ for an email and 15¢ for a phone. The most a run can cost is quoted before it starts, and a saved request reruns hourly to yearly.

OpenAI web search

A tool of the OpenAI API at $10 per 1,000 calls, plus search content tokens at the model's rates. SpecStream costs $5 per 1,000 searches with no tokens on top, and the results work with any model.

Claude web search

A tool of the Claude API at $10 per 1,000 searches, plus standard token costs for search-generated content. SpecStream returns the raw results at $5 per 1,000 for Claude or any other model.

Custom Search JSON API

Google's raw result list at $5 per 1,000 queries, closed to new customers and ending on January 1, 2027. SpecStream answers the same call in the same JSON. See the shutdown and the one-line move.

Brave Search API

A raw result list on Brave's own index at $5 per 1,000 requests, up to 50 queries a second, with $5 in free credits every month. SpecStream costs the same $5 and scores 97.9 F-score on all 1,000 SimpleQA Verified questions.