Brave Search API vs Tavily in 2026

Both calls on one key

A team does not have to pick one search API and rewrite the other. Each call keeps its path, its body and its parser; the address and the key change.

# Brave Search API, before and after
curl "https://api.search.brave.com/res/v1/web/search?q=open+source+vector+databases&count=10" \
  -H "X-Subscription-Token: $BRAVE_API_KEY"
curl "https://api.specstream.ai/res/v1/web/search?q=open+source+vector+databases&count=10" \
  -H "X-Subscription-Token: $SPECSTREAM_API_KEY"

# Tavily, before and after
curl -X POST https://api.tavily.com/search \
  -H "Authorization: Bearer $TAVILY_API_KEY" -H "Content-Type: application/json" \
  -d '{"query": "Who founded Mistral AI and when?", "max_results": 5}'
curl -X POST https://api.specstream.ai/tavily/search \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
  -d '{"query": "Who founded Mistral AI and when?", "max_results": 5}'

Each call of up to 10 results is one search at $0.005. A call that names no result count gets 10 results at that price, on Brave's path and on Tavily's.

What stays the same

On Brave's path, q, count from 1 to 20, offset and freshness are read, and country, search_lang and safesearch are accepted. The reply has type, query, mixed and web.results[], and a refusal keeps Brave's ErrorResponse shape. Each result carries one description, so extra_snippets comes back empty. Every parameter is in the Brave migration guide.

On /tavily, query, max_results, include_domains, exclude_domains, time_range and topic are read. The reply has query, answer, results[] with title, url, content, score and raw_content, response_time and request_id; favicon and images are not returned. Every parameter is in the Tavily migration guide.

$5 per 1,000, basic or advanced

Brave's Search plan costs $5 per 1,000 requests from prepaid credits. On Tavily pay as you go, 1,000 basic searches cost $8 and 1,000 advanced searches $16, and requests stop when the month's credits run out. SpecStream charges $5 per 1,000 searches of up to 10 results on both routes, with page text and the answer at no extra cost, from a balance that never expires.

The most accurate web search API, in about one second

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

Both routes run the fast search unless the call asks for page text, the answer or more than ten results. On November 9, 2026 production took 1 to 20 fast searches a second 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. There is no limit per key and no daily limit.

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, its answer is not reused for another call, and its log lines carry neither. 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 runs Deep Search, which returns every match for a request in plain words as structured rows at 5¢, 15¢ or $1 a row, with a source URL on every value. Save a request and it reruns hourly to yearly, with each row labeled changed, unchanged or no longer matching.

SpecStream

The live web data engine for teams and agents. Search at $5 per 1,000 on its own API or on the Brave, Tavily, Exa, Serper, Parallel and Firecrawl routes, pages at $0.80 per 1,000, and every match as structured rows at 5¢, 15¢ or $1 a row.

Exa

$4 per 1,000 requests on lite and $7 on auto and fast. SpecStream answers Exa's POST /search at https://api.specstream.ai/exa, in Exa's JSON, at $5 per 1,000.

Serper

Google results from prepaid packs, $1.00 per 1,000 on the 50k pack, with credits valid for 6 months. SpecStream answers Serper's call at https://api.specstream.ai/serper/search, and its credit never expires.

Parallel

Search from $1 per 1,000 requests. SpecStream answers Parallel's POST /v1/search at https://api.specstream.ai/parallel in Parallel's JSON, and the same key reads pages and returns rows.

Firecrawl search

2 credits per search, from monthly plans that start at $19. SpecStream answers Firecrawl's search at https://api.specstream.ai/firecrawl. Tavily vs Firecrawl.