Web search in n8n

The setting is one field

In n8n, add the SearXNG tool to your AI Agent node, create a SearXNG credential, and fill API URL. The credential has no key field, so the key goes in the address as its path part.

Credential: SearXNG
  API URL   https://api.specstream.ai/searxng/<key>

n8n strips a trailing slash, adds /search, and sends q, format=json, pageno and safesearch, with language when you set it. SpecStream answers with results[], each with title, url and content, the shape the tool formats for the agent.

Check the address before you save it

This is the call n8n's SearXNG tool sends. Run it with your key and you see what your agent will read.

curl -X POST "https://api.specstream.ai/searxng/$SPECSTREAM_KEY/search?q=open+source+vector+databases&format=json&pageno=1" \
  -H "Accept: application/json"

$5 per 1,000 searches, with nothing to subscribe to

Each search the agent makes is one search of up to 10 results at $0.005, taken from your balance. No monthly plan is needed and there is no daily limit, a top-up never expires, and a call stopped by a 429 charges nothing. An agent that searches 300 times a day costs $1.50 a day.

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, inside n8n's 15 second timeout

The SearXNG route runs on the fast search. On November 9, 2026 production took 1, 5, 10 and 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. n8n's SearXNG tool waits 15 seconds before it gives up.

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: not in the history, not in a log line. spend_cap_cents is the most the key can spend, so an agent in a loop stops at the cap you set.

curl -X POST https://api.specstream.ai/v1/account/keys \
  -H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
  -d '{"name": "n8n", "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: a request in plain words returns every match on the live web as rows, at 5¢, 15¢ or $1 a row, with the most a run can cost quoted before it starts. A saved request reruns on a schedule and sends its rows to n8n, Zapier, Make or a signed webhook.

Brave Search API

A search API on Brave's own index at $5 per 1,000 requests, with $5 in free credits every month and up to 50 queries per second. n8n's Brave Search credential takes only a key. On SpecStream the same $5 per 1,000 searches the whole live web, set in the SearXNG credential.

SerpApi

Search engine result pages as JSON, sold by monthly plan: Starter is $25 a month for 1,000 searches, with 200 successful searches an hour, and the free plan has 250 searches a month. On SpecStream $25 pays for 5,000 searches, and a top-up never expires.

Google Custom Search JSON API

$5 per 1,000 queries after 100 free a day, up to 10,000 a day, for existing customers only: Google closed it to new customers and ends it on January 1, 2027. SpecStream costs the same $5 per 1,000, takes new accounts today, and has no daily limit.

SearXNG on your own server

n8n's SearXNG tool is built for an instance you run and keep up yourself. SpecStream answers the same tool at the same field, with no server to run, and scores 97.9 F-score on SimpleQA Verified.