scrapeprices 
Jina Reader
Jina Reader is the option that looks cheapest and cannot be compared. Prepending r.jina.ai to a URL is genuinely free with no key at 20 requests a minute, and paid tokens run $0.050 per million falling to $0.045, but the meter counts the tokens in the output, so a long page costs several times what a short one does and nobody can tell you what 1,000 pages will cost until they have scraped them. That is why every vertical here is blank: turning tokens into pages would mean inventing an average page size, and the whole point of this table is not doing that. Two things inside the meter change the bill materially: the readerlm-v2 engine, the one you would pick for messy HTML, costs 3x tokens, and the search endpoint charges a flat 10,000 tokens minimum per request whatever comes back. What Jina does do better than most of this list is the failure rule, stated plainly: "tokens are not deducted for failed requests". Balances never expire, there is no subscription and no monthly minimum, and a country-specific proxy is a header away. There is no residential pool and no anti-bot tier, and the 10M free tokens are licensed for non-commercial use only. Jina AI GmbH is Berlin-based and has been an Elastic company since October 2025.
Plain page, by volume
| Pages per month | Monthly bill | Per 1,000 | Cheapest route | Allowance unused |
|---|---|---|---|---|
| 1k | cost not published | - | - | - |
| 10k | cost not published | - | - | - |
| 100k | cost not published | - | - | - |
| 1M | cost not published | - | - | - |
How Jina Reader compares
Jina Reader publishes no price at 10,000 pages a month: cost not published. The cheapest published route here costs $0.130 per 1,000 and AnyAPI costs $0.700.
| Provider | 1k a month | 10k a month | 100k a month | 1M a month |
|---|---|---|---|---|
Z  Zyte API | $0.130 | $0.130 | $0.130 | $0.100 |
A  AnyAPI | $0.700 | $0.700 | $0.700 | $0.700 |
N  Nimble | $1.00 | $1.00 | $1.00 | $1.00 |
| J Jina Reader | cost not published | cost not published | cost not published | cost not published |
List prices, dollars per 1,000 Plain pages at the cheapest published route to each volume. How this is worked out
Every published plan
| Plan | Price | Included credits | Overage each | Note |
|---|---|---|---|---|
| Toy Experiment (free key) | $0 one-time, never expires | 10,000,000 | hard cap | 10 million tokens on every new API key, one time, not monthly. "Non-commercial use only (CC-BY-NC)." 500 RPM on r.jina.ai. includedUnits is tokens, not pages. |
| Prototype Development | $50.00 one-time, never expires | 1,000,000,000 | hard cap | $50 buys 1 billion tokens, which the widget prints as "0.050 / 1M tokens". Standard key, 500 RPM on r.jina.ai. A top-up, not a subscription, and the widget shows no expiry. includedUnits is tokens, not pages. |
| Production Deployment | $500 one-time, never expires | 11,000,000,000 | hard cap | $500 buys 11 billion tokens, printed as "0.045 / 1M tokens". This is the premium key: 5,000 RPM and 50M TPM on r.jina.ai, advanced key management, 24-hour support, one hour of integration consultation. includedUnits is tokens, not pages. |
Every vertical is null because Jina does not sell a page. The plan's unit is an output token, not a credit and not a request: the rate-limit table says the Reader API at r.jina.ai will "Count the number of tokens in the output response". A token cannot be turned into a page without inventing an average page length, so under this site's no-conversion rule the native figures go here instead. They are $50 for 1,000,000,000 tokens ($0.050 per 1M) and $500 for 11,000,000,000 tokens ($0.045 per 1M). Two real multipliers exist inside that token meter and are worth knowing: setting X-Engine to readerlm-v2 for HTML-to-Markdown conversion "Costs 3x tokens", and the search endpoint s.jina.ai is not metered on output at all but on "a fixed number of tokens, starting from 10000 tokens" per request. X-Token-Budget caps the tokens a single request may spend, and the request fails rather than overspending. Jina sells the plain fetch (X-Engine: direct), the browser render (the default engine) and Markdown; it sells no residential pool and no anti-bot tier, so those two are null for a second reason and are absent from the vertical list.
Cost by kind of fetch
Per 1,000, at 10,000 pages a month.
What it can do
Only what the provider's own pages state.
Where these numbers came from
- https://jina.ai/reader read 2026-09-14 / product and pricing page: the API Pricing widget, the rate-limit table ("Count the number of tokens in the output response", 20 RPM keyless, 500 RPM free and paid, 5000 RPM premium, 7.9s average latency), "Don't panic! Every new API key comes with 10M free tokens!", the FAQ answers on cost, on failed requests ("No, tokens are not deducted for failed requests"), on key expiry, and the X-Engine, X-Token-Budget, X-Proxy and X-Proxy-Url headers
- https://jina.ai/reader#pricing read 2026-09-14 / the three published packages read out of the widget: 10 Million tokens Free, non-commercial use only; 1 Billion tokens $50 at 0.050 / 1M tokens; 11 Billion tokens $500 at 0.045 / 1M tokens. No separate Premium price is listed, Premium is what the $500 package includes
- https://jina.ai/api-dashboard/ read 2026-09-14 / parameter panel: X-Engine readerlm-v2 "Uses ReaderLM-v2 for HTML-to-Markdown conversion ... Costs 3x tokens!", and X-Token-Budget "Limits the maximum number of tokens used for this request"
- https://jina.ai/about-us read 2026-09-14 / "Founded in 2020 by Han Xiao"; "On October 9, 2025, Jina AI was acquired by Elastic (NYSE: ESTC)"
- https://jina.ai/legal read 2026-09-14 / operating company and address: "Jina AI GmbH, Prinzessinnenstrasse 19-20, 10969 Berlin, Germany", with the note that Elastic now handles privacy enquiries
Something out of date? Send the provider's own page to hello@scrapeprices.com and it is corrected on the next capture.