Skip to the page

Seltz alternatives in 2026

Search from $1 per 1,000, pages at $0.80.

Seltz charges $3 per 1,000 searches on Base and $5 on Pro, its default. Fetch costs $5 to $10 per 1,000 URLs and Answer $6 to $7 per 1,000. Every result comes from Seltz's own index, which refreshes on its own schedule.

SpecStream is the live web data engine for teams and agents. Its Search API reads the live web when your call arrives, from $1 per 1,000 searches, and scores 97.9 on SimpleQA Verified. The Extract API reads pages at $0.80 per 1,000.

New accounts get $10 of credit, no card. It covers 10,000 searches.

SpecStream and Seltz side by side

SpecStreamSeltz
What you sendA question in everyday words, in any language, up to 2,000 characters: GET https://api.specstream.ai/v1/web-search?q=<the question> with your key as a Bearer token. POST takes the same parameters in a JSON bodyPOST https://api.seltz.ai/v1/search with query, max_results, a scope and a tier, and the key in the x-api-key header
What comes backanswer with the value, the page it was read on and the time it was read. rows with each result and the facts the request asks for. sources in rank order, each with its url, title, snippet and published datedocuments[], each with url and published_date, the page content up to 20,000 characters by default, and snippets or properties when fields asks for them. Your model reads them to find the answer
Price per 1,000 searches$1 for up to 10 results with speed=fast. $5 on the default search, which adds the answer with its page and the rows$3 on the Base tier and $5 on the Pro tier. A search with no tier runs on Pro
Results in one call1 to 100 results in one search, set with num, and page (1 to 10) continues the same ranked list. Up to 10 results cost $0.001 on the fast search, and each started block of 10 more adds $0.0012, so 100 results cost $0.0118max_results defaults to 10. A search costs the same however many documents it returns
Where the results come fromThe whole live web, read at the moment of the call. News and image searches run on the fast tier at the same priceSeltz's own index, not a live read. Its docs say the corpus refreshes on its own cadence, so a window of an hour or two can return nothing
Search accuracy97.9 F-score on all 1,000 official SimpleQA Verified questions (978 correct). 64.0 item F1 and 43.4 row F1 on all 200 official WideSearch tasks. Scores and method on the benchmarks pageSeltz's own pricing page says it topped Exa's People and Company Search benchmarks
Reading pagesExtract API: $0.80 per 1,000 pages, up to 20 URLs a call, PDF, Word and Excel included. A page with no text is not charged. A page that refuses a plain read is tried again in a real browserFetch: $5 per 1,000 URLs on Base and $10 per 1,000 on Pro, which renders pages that need a browser. Up to 20 URLs a call
An answer with its sourceInside the $5 default search: answer with the value, the page it was read on and its read time, and every source that prints itAnswer: $6 per 1,000 on Base and $7 per 1,000 on Pro
Research and datasetsDeep Search: a request in plain words returns every match on the live web as rows that meet each condition you wrote, 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 nothing runs until you approve itAgent: $0.25 a run on low effort, $0.50 on medium (the default), $1.00 on high and $2.00 on max, where a run can take 30 minutes or more
Reruns and change labelsA saved request reruns hourly to yearly. Each row comes back labeled changed, unchanged or no longer matching. A changed or new row costs half its rate, an unchanged row 1¢Monitor reruns a saved search on a cadence and hands back what is new: $3 per 1,000 checks on Base, $5 on Pro
Free to start$10 of credit, no card. That is 10,000 fast searches, 2,000 default searches or 12,500 pages$20 of free requests on a new account, then pay as you go once you add a payment method. Qualified startups can apply for $1,000 in credits
Data already heldPublic databases read every morning at specstream.ai/data: every company that filed a Form D since 2023 with its filing history, nearly 2 million US nonprofits, every US bank failure since 1934, federal rules, open federal contracts and grants, and local businesses. Signals such as funding rounds, new executives and layoffs at specstream.ai/signalsSearch scopes for web (in beta), news, Wikipedia, people, companies and financial reports
Other search APIs on the same keyCalls written for Tavily, Exa, Serper, Brave, Perplexity, You.com, Linkup, Parallel, Bing v7, Google Custom Search, Firecrawl and SearXNG are answered in their own JSON at api.specstream.ai, at $1 per 1,000 searches. Change the base address and the keySearch, Answer, Monitor, Fetch and Agent on Seltz's own API
MCP, SDKs and where the data goesMCP server at https://api.specstream.ai/mcp, the REST API, Python and TypeScript SDKs, a CLI and A2A. Rows go to Google Sheets, Excel, Airtable, Notion, HubSpot, Salesforce, Pipedrive, Attio, Slack, Teams, Zapier, Make, n8n and signed webhooksMCP server at https://mcp.seltz.ai/mcp, the REST API, and Python and JavaScript SDKs

Why teams move from Seltz to SpecStream

Other Seltz alternatives

How SpecStream scores

The same numbers as the benchmarks page, from the official tasks and scorers.

How to switch from Seltz

1. Create a key under Agents & API in the workspace. A new account gets $10 of credit, no card, which covers 10,000 fast searches. Give the key a spending limit.

2. Replace POST https://api.seltz.ai/v1/search with GET https://api.specstream.ai/v1/web-search. query becomes q, max_results becomes num, and the key goes as a Bearer token in place of x-api-key. Add speed=fast for $1 per 1,000 searches, and passages=true for the text of each page.

import requests

reply = requests.get(
    "https://api.specstream.ai/v1/web-search",  # was POST https://api.seltz.ai/v1/search
    headers={"Authorization": f"Bearer {SPECSTREAM_KEY}"},  # was x-api-key
    params={"q": "Who is the CEO of Mistral AI", "num": 10, "speed": "fast", "passages": "true"},
).json()["data"]

for source in reply["sources"]:  # was documents
    print(source["url"], source.get("published"), source.get("excerpt") or source["snippet"])

3. Map the fields. documents[].url is sources[].url, published_date is published, and content is excerpt, up to about 3,500 characters of the page closest to the question. include_domains becomes site: in q, and from_date becomes after: in q. Drop tier and scope: the default search, with no speed, returns the answer with its page at $5 per 1,000.

4. Move Fetch to the Extract API: POST https://api.specstream.ai/v1/extract with {"urls": ["<page>"]}, up to 20 URLs a call. Each result in data.results has its url, status, title and text, at $0.80 per 1,000 pages.

5. Connect an agent through the MCP server at https://api.specstream.ai/mcp in place of https://mcp.seltz.ai/mcp. Its search_web tool runs a search for $0.005, from the same balance and inside the limits of its key. Setup for each client is in the MCP docs.

Frequently asked questions

Something we didn't answer? Talk to a human

How much does Seltz cost?
Search costs $3 per 1,000 on Base and $5 on Pro, the default. Fetch costs $5 or $10 per 1,000 URLs, Answer $6 or $7 per 1,000, Monitor $3 or $5 per 1,000 checks, and an Agent run $0.25 to $2.00 by effort. New accounts get $20 of free requests.
What is the best Seltz alternative?
SpecStream, from $1 per 1,000 searches. It searches the live web at the moment of the call, returns the answer with the page it was read on for $5 per 1,000, reads pages at $0.80 per 1,000, and scores 97.9 F-score on all 1,000 SimpleQA Verified questions. New accounts get $10 of credit, no card.
Does Seltz search the live web?
No. Seltz serves results from its own index, and its docs say the corpus refreshes on its own cadence. SpecStream searches the whole live web when the call arrives, for $1 per 1,000 searches with speed=fast.
What is the cheapest alternative to Seltz Fetch?
The SpecStream Extract API, at $0.80 per 1,000 pages against $5 to $10 per 1,000 URLs on Fetch. It takes up to 20 URLs a call, reads PDF, Word and Excel files, and does not charge a page with no text.
How accurate is the SpecStream Search API?
It scores a 97.9 F-score on all 1,000 official SimpleQA Verified questions, with 978 correct, and 64.0 item F1 on all 200 official WideSearch tasks. The scores and the method are on the benchmarks page.
Does SpecStream have an MCP server?
Yes, at https://api.specstream.ai/mcp over Streamable HTTP, with OAuth or an API key. The search_web tool runs a search for $0.005, and the agent spends from the same balance as your team, inside the limits you set on its key.

Sources

SpecStream wrote this page. Each rival figure comes from that company's own page.

  1. Seltz pricing (docs)docs.seltz.ai/pricing.md, read Nov 12, 2026
  2. Seltz pricing pageseltz.ai/pricing, read Nov 12, 2026
  3. Seltz Search API referencedocs.seltz.ai/api-reference/search.md, read Nov 12, 2026
  4. Exa pricingexa.ai/docs/admin/pricing.md, read Nov 12, 2026
  5. Tavily credits and pricingdocs.tavily.com/documentation/api-credits.md, read Nov 12, 2026
  6. Brave Search APIbrave.com/search/api/, read Nov 12, 2026
  7. Linkup pricingdocs.linkup.so/pages/documentation/platform/pricing.md, read Nov 12, 2026
  8. SpecStream benchmarksspecstream.ai/benchmarks, read Nov 12, 2026
  9. SpecStream pricingspecstream.ai/pricing, read Nov 12, 2026
  10. SpecStream for agentsspecstream.ai/agents.md, read Nov 12, 2026
  11. SpecStream dataspecstream.ai/data, read Nov 12, 2026