Firecrawl migration
Change the base address, keep search and scrape.
Since November 9, 2026 SpecStream answers Firecrawl v2 search and scrape at https://api.specstream.ai/firecrawl, in Firecrawl's JSON. A search costs $0.005 for up to 10 results and a scrape $0.0008 a page, from a balance, with no monthly plan.
Change https://api.firecrawl.dev to https://api.specstream.ai/firecrawl and send a SpecStream key as the Bearer token. POST /v2/search and POST /v2/scrape keep their bodies.
97.9 F-score on all 1,000 SimpleQA Verified questions for search, and 0.908 F1 on WCXB for the Extract API that answers each scrape, on the benchmarks page.
New accounts get $10 of credit, no card. It covers 2,000 searches.
Firecrawl v2 to SpecStream, parameter by parameter and field by field
| SpecStream (`/firecrawl/v2`) | Firecrawl | |
|---|---|---|
| Address | https://api.specstream.ai/firecrawl, so the calls are POST /firecrawl/v2/search and POST /firecrawl/v2/scrape. A setting that holds the version takes https://api.specstream.ai/firecrawl/v2 | https://api.firecrawl.dev, with POST /v2/search and POST /v2/scrape |
| Key | Your SpecStream key as Authorization: Bearer, or as the HTTP Basic password. No key gets a 401 | Your Firecrawl key as Authorization: Bearer |
| Search: `query` and `limit` | Read. limit is 1 to 100, default 5. Up to 10 results cost $0.005 and run the fast search; 100 cost $0.059 | The search body with query and limit. A search takes 2 credits |
| Search: `includeDomains`, `excludeDomains`, `tbs` | Read as site filters and as a date filter (qdr: or cdr with cd_min and cd_max) | Fields of the search body |
| Search: `scrapeOptions.formats` with `markdown` | Read. Each result then carries markdown: the text of its page closest to the query, up to about 3,500 characters, and metadata. Same price; it runs the default search | Asks for each result's page as markdown, as LobeHub sends it with a limit of 20 |
| Search: `timeout`, `ignoreInvalidURLs` and other fields | Accepted with no error and not used | Fields Open WebUI sends with its search |
| Search reply | success: true, data.web[] with url, title, description and position, id, and specstream with the price of the call. data.web is the only list in data | success: true and data.web[] with url, title and description, and markdown when asked |
| Scrape: `url` | Read. One page or document (PDF, Word, Excel) through the Extract API. Other formats are accepted; the reply is markdown | The page to scrape, with the formats you want back |
| Scrape reply | success: true, data.markdown (the main text, without menus, footers and ads) and data.metadata with sourceURL, url, statusCode, title and publishedTime when the page gives one. data.html and data.rawHtml are not returned | data.markdown, data.html, data.rawHtml and data.metadata |
| A page that gives no text | success: false with a 500 and the reason, and nothing is charged. A page that refuses a plain read is tried as a browser would ask, then in a real browser, then from the site's feed, then from its newest archive capture | success: false with an error |
| Price | $0.005 a search of up to 10 results and $0.0008 a page, paid from a balance. Credit you top up never expires | Monthly plans in credits: Hobby $19 a month for 5,000, Standard $99 a month for 100,000, Scale $749 a month for 1,000,000 |
| Before you pay | $10 of credit on sign up, no card: 2,000 searches or 12,500 pages | A free plan of 500 searches or 1,000 pages scraped, at 10 requests a minute |
What changes and what stays when you move from Firecrawl
The whole migration is the base address and the key
Your code keeps Firecrawl's paths, bodies and parser. The base address changes, and the Bearer key.
# Before
curl -X POST https://api.firecrawl.dev/v2/search \
-H "Authorization: Bearer $FIRECRAWL_API_KEY" -H "Content-Type: application/json" \
-d '{"query": "open source vector databases", "limit": 5}'
# After
curl -X POST https://api.specstream.ai/firecrawl/v2/search \
-H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
-d '{"query": "open source vector databases", "limit": 5}'
# A scrape is one Extract API page, $0.0008
curl -X POST https://api.specstream.ai/firecrawl/v2/scrape \
-H "Authorization: Bearer $SPECSTREAM_API_KEY" -H "Content-Type: application/json" \
-d '{"url": "https://example.com/pricing"}'Firecrawl's JavaScript SDK works with its API address set to https://api.specstream.ai/firecrawl, and LobeHub takes https://api.specstream.ai/firecrawl/v2. The Python SDK, firecrawl-py, works with api_url="https://api.specstream.ai", and with it the Firecrawl loaders of GPT Researcher, LangChain and LlamaIndex.
What the reply leaves out
The route answers search and scrape. A search returns data.web only. markdown on a search result is the part of the page closest to the query, up to about 3,500 characters; for the whole page, scrape it. A scrape returns markdown and metadata, so data.html and data.rawHtml are not in the reply.
A refusal keeps Firecrawl's shape: success: false with error and a code. A missing key is a 401, an empty balance a 402, and a busy minute a 429 with Retry-After that charges nothing.
Pay per call, with no plan to buy first
Firecrawl sells credits by month: Hobby is $19 a month for 5,000 credits and Standard $99 a month for 100,000, and a search takes 2 credits. SpecStream takes $0.005 a search and $0.0008 a page from your balance, a page with no text is not charged, and the credit you top up never expires. Every rate is on the pricing page.
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, measured under load
Every drop-in route runs the fast search unless the call asks for page text or results past the first ten. 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. One account takes up to 8,800 searches a minute; a call over that gets a 429 with Retry-After and is not charged.
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 calls GET /v1/web-search, which returns up to 100 results for one search and an answer with the page it was read from. It 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, and the most a run can cost is approved before it starts.
Other options when you leave Firecrawl
Exa
A search API from $4 per 1,000 requests, with page contents at $1 per 1,000 pages. SpecStream reads pages at $0.80 per 1,000 and answers Exa's own call at https://api.specstream.ai/exa.
Tavily
Search and extract billed in credits at $0.008 a credit on pay as you go, with 1,000 free credits a month. SpecStream answers Tavily's call at https://api.specstream.ai/tavily at $0.005 a search.
Brave Search API
A search API on Brave's own index at $5 per 1,000 requests, with $5 in free credits every month. It returns search results; SpecStream returns the results and reads the pages behind them on the same key.
Serper
Google results as JSON, in credit packs from $1.00 per 1,000 that are valid for 6 months. SpecStream answers Serper's call at https://api.specstream.ai/serper, reads the pages too, and its top-up credit never expires.
SerpApi
Search engine result pages as JSON, sold by monthly plan: Starter is $25 a month for 1,000 searches. On SpecStream $25 pays for 5,000 searches or 31,250 pages.
How SpecStream scores
The same numbers as the benchmarks page, from the official tasks and scorers.
| SpecStream |  Parallel |  Exa |  Tavily |  Perplexity | |
|---|---|---|---|---|---|
| Web researchWideSearch, GPT-5.6 Luna | 60.9 | 45.5 | 53.0 | 47.9 | 47.0 |
| Factual accuracySimpleQA Verified, GPT-5.6 Luna | 96 | 94 | 91 | 94 | 94 |
| Factual accuracySimpleQA Verified, GPT-5.6 Sol | 99 | 97 | 91 | 92 | 95 |
How to move from Firecrawl
1. Create a key under Agents & API in the workspace. A new account gets $10 of credit, no card, which covers 2,000 searches.
2. Set the base address to https://api.specstream.ai/firecrawl where your code has https://api.firecrawl.dev, and send the SpecStream key as the Bearer token. In Python:
import os
import requests
BASE = "https://api.specstream.ai/firecrawl" # was https://api.firecrawl.dev
HEADERS = {"Authorization": f"Bearer {os.environ['SPECSTREAM_API_KEY']}"}
found = requests.post(
f"{BASE}/v2/search",
json={"query": "open source vector databases", "limit": 5},
headers=HEADERS,
).json()
for result in found["data"]["web"]:
print(result["title"], result["url"], result["description"])
page = requests.post(
f"{BASE}/v2/scrape",
json={"url": found["data"]["web"][0]["url"]},
headers=HEADERS,
).json()
print(page["data"]["metadata"].get("title"), page["data"]["markdown"][:500])In JavaScript:
const BASE = "https://api.specstream.ai/firecrawl"; // was https://api.firecrawl.dev
const headers = {
Authorization: `Bearer ${process.env.SPECSTREAM_API_KEY}`,
"Content-Type": "application/json",
};
const found = await fetch(`${BASE}/v2/search`, {
method: "POST",
headers,
body: JSON.stringify({ query: "open source vector databases", limit: 5 }),
}).then((response) => response.json());
for (const result of found.data.web) {
console.log(result.title, result.url, result.description);
}3. Keep limit, includeDomains, excludeDomains and tbs as they are. Add scrapeOptions with formats: ["markdown"] when you want the part of each page closest to the query in the search reply.
4. Read data.markdown from a scrape, and treat data.html and data.rawHtml as optional. Each reply carries the price of the call under specstream.
Frequently asked questions
Something we didn't answer? Talk to a human
What is the best Firecrawl alternative?
https://api.specstream.ai/firecrawl in Firecrawl's JSON, and its Extract API scores 0.908 F1 on WCXB.How do I migrate from Firecrawl?
https://api.firecrawl.dev with https://api.specstream.ai/firecrawl and send a SpecStream key as the Bearer token. POST /v2/search and POST /v2/scrape keep their bodies.Is there a cheaper Firecrawl alternative?
Is there a free Firecrawl alternative?
Does search with `scrapeOptions` return the whole page?
/v2/scrape for $0.0008.Does the Firecrawl SDK work with SpecStream?
https://api.specstream.ai/firecrawl. The Python SDK works with api_url="https://api.specstream.ai", and so do the Firecrawl loaders of GPT Researcher, LangChain and LlamaIndex.Is there a Firecrawl alternative for Claude?
https://api.specstream.ai/mcp, and its read_pages tool reads up to 20 pages or documents a call at $0.80 per 1,000 pages.Sources
SpecStream wrote this page. The route's behavior is read from its code; each rival figure comes from that company's own page.
- Firecrawl pricingfirecrawl.dev/pricing, read Nov 9, 2026
- Exa pricingexa.ai/docs/admin/pricing.md, read Nov 9, 2026
- Tavily pricingtavily.com/pricing, read Nov 9, 2026
- Brave Search APIbrave.com/search/api/, read Nov 9, 2026
- Serperserper.dev/, read Nov 9, 2026
- SerpApi pricingserpapi.com/pricing, read Nov 9, 2026
- SpecStream benchmarksspecstream.ai/benchmarks, read Nov 10, 2026
- SpecStream pricingspecstream.ai/pricing, read Nov 10, 2026
- SpecStream for agentsspecstream.ai/agents.md, read Nov 10, 2026