Metaso Search API compatible

The call, before and after

# Before
curl -X POST https://metaso.cn/api/v1/search \
  -H "Authorization: Bearer $METASO_API_KEY" -H "Content-Type: application/json" \
  -d '{"q": "Who founded Mistral AI and when?", "scope": "webpage", "size": 10, "includeSummary": true}'

# After
curl -X POST https://api.specstream.ai/metaso/api/v1/search \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
  -d '{"q": "Who founded Mistral AI and when?", "scope": "webpage", "size": 10, "includeSummary": true}'

The body and the parser stay. The base address changes to https://api.specstream.ai/metaso/api/v1, and the key changes. Every parameter and field below is what the route answers on November 12, 2026.

The MCP server that takes the address as a setting

metaso-search-mcp, the community MCP server for Metaso's API, reads its base address from METASO_BASE_URL. Set it to SpecStream's address and put your SpecStream key in METASO_API_KEY:

{
  "mcpServers": {
    "metaso-search-mcp": {
      "command": "npx",
      "args": ["metaso-search-mcp"],
      "env": {
        "METASO_API_KEY": "YOUR_SPECSTREAM_KEY",
        "METASO_BASE_URL": "https://api.specstream.ai/metaso/api/v1"
      }
    }
  }
}

Its metaso_search and metaso_reader tools then call SpecStream. Its paper, video and podcast scopes answer 400.

Every parameter the route uses

Used in the search: q, scope (webpage by default, document for PDF files, image for images), size (1 to 100, sent as a number or a string) and page (from 1, each page continuing the same ranked list).

Used to shape the reply: includeSummary or includeRawContent set to true reads each result's page and fills summary or rawContent with the text of that page closest to the query, up to about 3,500 characters. A page not read in time keeps its snippet. For a whole page, send its URL to the reader. conciseSnippet comes back in searchParameters.

Every field that comes back

The search reply has searchParameters (q, scope, size, includeSummary, includeRawContent, conciseSnippet), total, and the list for the scope: webpages, documents or images. Each item has title, link, snippet and position, date when the page states one, and summary or rawContent when you ask for them. An image item adds imageUrl and thumbnail.

Not returned: credits, Metaso's count of the credits a call used.

One field is added: specstream, with what the Search API returns beyond Metaso's fields, including the call's price_usd. A parser that reads Metaso's fields never sees it.

The reader

curl -X POST https://api.specstream.ai/metaso/api/v1/reader \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
  -H "Accept: application/json" \
  -d '{"url": "https://www.python.org/downloads/"}'

With Accept: text/plain the reply is the page as Markdown. With Accept: application/json it is title, url, date when the page states one, and markdown. Each page read is one Extract API page at $0.0008. A page that cannot be read answers 422 and is not charged.

Errors in Metaso's form

A refused key keeps its HTTP status, 401, and answers in Metaso's body: errCode 2005 with errMsg, plus message, which metaso-search-mcp prints. A scope this route does not search answers 400. When search is busy, the route answers 429 with Retry-After, and nothing is charged.

$1 per 1,000 searches, page text included

A search of up to 10 results costs $0.001, with or without includeSummary and includeRawContent. Each started block of 10 more results adds $0.0012: size 20 costs $0.0022, size 50 costs $0.0058 and size 100 costs $0.0118. A reader call costs $0.0008 a page. A call stopped by a 429 charges nothing.

The most accurate web search API

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.

You decide what is kept, on each key

Create the 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 a page with no text is not charged. Deep Search 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. The most a run can cost is approved before it starts.

Exa

A search API on Exa's own index, at $7 per 1,000 requests on auto and $4 on lite for up to 10 results, with its own request and reply. SpecStream answers Exa's call as well, at https://api.specstream.ai/exa, for $0.001 a search.

Tavily

A search API sold in credits at $0.008 each pay as you go: a basic search takes 1 credit and an advanced search 2. SpecStream answers Tavily's POST /search at https://api.specstream.ai/tavily for $0.001 a basic search.

Brave Search API

A search API on Brave's own index at $5 per 1,000 requests and 50 queries per second. SpecStream answers Brave's call at https://api.specstream.ai/res/v1/web/search at $1 per 1,000, a fifth of Brave's rate.