Web search APIs for company news
Query in, ranked web results out. These endpoints search the open web rather than a curated index, and are ranked here against each other with the news indexes held out, so the comparison is between substitutes.
Measured on the same questions, the strongest dedicated news index, PredictLeads, reaches 69.0%. Every endpoint below beats it, which is why the two are ranked apart.
News index APIs for company news · Most accurate · Fastest · Cheapest · Full company news board · Web search API benchmark
Web search APIs for company news, all metrics
10 endpoints, ranked by extracted-answer accuracy on 129 questions.
| Rank | Vendor | Endpoint & configuration | Accuracy | AR@1 | AR@5 | Latency | Snippet tokens | Total $ | Official docs |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Exaweb search | POST /searchtype=deep | 99.2% | 98.5% | 100.0% | 3.00s | 1,330 | $1.65 | Official docs |
| 2 | Exaweb search | POST /searchtype=instant | 97.7% | 83.7% | 97.7% | 447ms | 2,051 | $1.05 | Official docs |
| 3 | Parallelweb search | POST /v1/searchmode=advanced | 96.9% | 62.0% | 96.9% | 3.28s | 2,127 | $0.80 | Official docs |
| 4 | Linkupweb search | POST /v1/searchdepth=standard · outputType=searchResults | 96.1% | 80.6% | 93.0% | 2.34s | 2,926 | $0.83 | Official docs |
| 5 | Parallelweb search | POST /v1/searchmode=basic | 95.3% | 65.1% | 91.5% | 1.41s | 2,409 | $0.82 | Official docs |
| 6 | Linkupweb search | POST /v1/searchdepth=fast · outputType=searchResults | 94.6% | 88.4% | 95.3% | 1.82s | 2,838 | $0.82 | Official docs |
| 7 | Tavilyweb search | POST /searchsearch_depth=advanced · chunks_per_source=3 | 93.8% | 79.8% | 94.6% | 4.40s | 3,041 | $2.25 | Official docs |
| 8 | Brave Searchweb search | GET /res/v1/web/searchcount=10 · result_filter=web | 93.8% | 83.7% | 93.8% | 693ms | 723 | $0.73 | Official docs |
| 9 | SERPgoogle search | GET google-search74.p.rapidapi.comlimit=10 | 93.0% | 73.6% | 93.0% | 2.04s | 469 | $0.46 | Official docs |
| 10 | Firecrawlweb search | POST /v2/search | 92.3% | 74.4% | 92.3% | 1.42s | 672 | $0.73 | Official docs |
News index APIs for company news, same questions
Measured on the identical question set with the same judge, ranked separately because it is a different mechanism.
| Rank | Vendor | Endpoint & configuration | Accuracy | AR@1 | AR@5 | Latency | Snippet tokens | Total $ | Official docs |
|---|---|---|---|---|---|---|---|---|---|
| 1 | PredictLeadscompany news events | GET /api/v3/companies/{domain}/news_eventscategories[] · paginated | 69.0% | 62.0% | 72.1% | 651ms | 302 | $5.23 | Official docs |
| 2 | Seltznews index | POST /v1/searchscope=news | 66.7% | 31.8% | 63.6% | 330ms | 2,915 | $0.82 | Official docs |
Web search APIs for company news
What counts as a general web search API here?
An endpoint that takes a query and searches the open web, returning ranked results the caller reads. Curated news indexes are measured on the same questions but ranked separately, because they are a different mechanism.
How are these endpoints made comparable?
Every endpoint receives the identical user query, may return at most 10 results, and is scored by one judge against ground truth pinned to official sources. Nothing is rewritten per vendor. 129 questions, one run.
Why do some vendors appear more than once?
Several expose more than one search depth or mode, and those behave differently enough on speed and cost to be worth separating. Each endpoint is its own row, with its request configuration shown beside it.