Best Web Scraping API for Indie Developers (2026)
For most indie developers, the best web scraping API is the one whose credit multipliers match your targets. Firecrawl is the easiest way to turn pages into LLM-ready markdown, ZenRows and Scrapfly are the cheapest entry points for general scraping, Zyte bills per successful response, and open source Crawl4AI, Scrapy or Crawlee cost nothing but your server and your time.
Key takeaways
- Headline credit counts can mislead: JS rendering costs 5 credits on ScrapingBee and ZenRows, adds 5 credits on Scrapfly and costs 10 credits on ScraperAPI, so the same plan buys far fewer rendered pages.
- Firecrawl charges 1 credit per scraped page and lists no separate JS rendering charge, but its credits are pricier: Hobby is $19/mo for 5,000 credits, against ScrapingBee's $19/mo for 75,000.
- Zyte API is pay as you go: $0.13 per 1,000 HTTP responses and $1.01 per 1,000 browser-rendered responses at its simplest tier, with $5 of free credit to start.
- Free tiers good enough to prototype on: ZenRows 5,000 credits a month, Bright Data Web Unlocker 5K requests a month, Firecrawl 1,000 credits a month, and Jina Reader 10M tokens per new API key.
- robots.txt is not access authorization under RFC 9309, and US cases such as hiQ v. LinkedIn and Meta v. Bright Data turned on narrow facts. Scraping personal data brings in the GDPR. None of this is legal advice.
On this page 15
- TL;DR
- What a scraping API actually sells you
- Comparison table
- What the table does not show
- Decision table by job
- Worked cost example: 100,000 pages a month
- Legality and ethics
- Frequently asked questions
- What is the cheapest web scraping API for an indie developer?
- Which scraping API is best for RAG and AI apps?
- How much does JavaScript rendering cost on scraping APIs?
- Is web scraping legal?
- Should I use an open source scraper instead of an API?
- Launch what you build
- Sources
Short answer: For feeding an AI app or RAG pipeline, start with Firecrawl (1,000 free credits a month, clean markdown out). For general scraping on a budget, ZenRows or Scrapfly give the most credits per dollar at entry level, and Zyte API is the cheapest pay-as-you-go option for plain HTML. If your targets are friendly and you run a server, open source Scrapy, Crawlee or Crawl4AI cost nothing per page.
Every price and multiplier below comes from the vendor's own page, checked on 13 Nov 2026. Advice is ours and marked as such.
TL;DR#
- Read the multipliers, not the headline credits. JS rendering costs 5 credits on ScrapingBee and ZenRows and 10 on ScraperAPI (ScraperAPI credits docs).
- Best for AI apps: Firecrawl, 1 credit per page, from $19/mo for 5,000 credits (Firecrawl pricing).
- Cheapest entry plans: ZenRows $19/mo for 45,000 credits, Scrapfly $30/mo for 200,000.
- Pay per success: Zyte API, $0.13 per 1,000 HTTP responses at its simplest tier (Zyte pricing).
- Free route: Scrapy, Crawlee or Crawl4AI on your own box.
Disclosure: ZeroRaised publishes this guide and is a launch site for bootstrapped products. ZeroRaised has no commercial relationship with any vendor listed here and earns nothing if you sign up. Facts come from the linked sources as checked on 13 Nov 2026.
What a scraping API actually sells you#
Fetching one page is easy. Fetching thousands an hour from sites that do not want bots is not. A scraping API bundles rotating proxies (datacenter or pricier residential), a headless browser for JavaScript-heavy pages, and unblocking (retries, fingerprints, CAPTCHA handling). Each costs the vendor money, so almost every API uses credit multipliers: a plain HTML fetch is 1 credit, while JS rendering or premium proxies can make the same request cost 5, 10, 25 or 30. Our read: work out your mix of static and rendered pages before you compare plans.
Comparison table#
Entry price is the cheapest paid plan on monthly billing. "JS rendering cost" is the vendor's own figure for a request with a headless browser.
| Tool | Entry price | Free credits | JS rendering cost | Best for |
|---|---|---|---|---|
| Firecrawl | Hobby $19/mo, 5,000 credits | 1,000 credits/mo, no card | No separate charge listed; 1 credit/page base, JSON extraction +4 | LLM-ready markdown, RAG |
| ScrapingBee | Hobby $19/mo, 75,000 credits | 1,000 trial credits, no card | 5 credits; premium proxy 10 or 25 | General scraping, simple API |
| ZenRows | Build $19/mo, 45,000 credits | 5,000 credits/mo, no card | 5x; premium 10x; both 25x | Cheap start, protected sites |
| ScraperAPI | Hobby $49/mo, 100,000 credits | 7-day trial, 5,000 credits | 10 credits; with premium 25 | Domain-specific endpoints |
| Scrapfly | Discovery $30/mo, 200,000 credits | 1,000 credits on signup | +5 credits (6 total on datacenter) | Most base credits per dollar |
| Zyte API | Pay as you go | $5 credit, 30 days | $1.01 vs $0.13 per 1,000 (tier 1) | Pay per success, Scrapy users |
| Bright Data Web Unlocker | $1.5 per 1,000 requests | 5K requests/mo free tier | No separate charge listed | Hard targets at volume |
| Oxylabs Web Scraper API | Starter $15/mo | $5 welcome credit | $1.50 vs $1.00 per 1,000 (custom domains, JS vs non-JS) | Enterprise-grade unblocking |
| Apify | Starter $19/mo of platform usage | $5/mo platform usage | Compute units, $0.20/CU on Free and Starter | Ready-made scrapers (Actors) |
| Browserbase | Developer $20/mo, 100 browser hours | 1 browser hour | Billed by browser hour ($0.12 overage on Developer) | Agents, logins, interactive flows |
| Jina Reader | Token-based | 10M tokens per new API key | Default engine renders JS | Quick URL-to-markdown |
Notes: ScraperAPI charges more for some domains (5 credits for Amazon, 25 for Google, 30 for LinkedIn), and its FAQ also mentions a 1,000-credit free plan. Scrapfly's residential pool costs 25 credits, or 30 with rendering (Scrapfly billing docs). Jina showed no per-token price when we checked.
What the table does not show#
- Firecrawl sells cleaning, not raw fetches. Its billing docs say a scrape is "1 credit / page", the enhanced proxy adds nothing and JSON extraction adds 4 credits (Firecrawl billing). Our read: per credit it is pricey next to ScrapingBee or Scrapfly, but if you would otherwise strip boilerplate before chunking for retrieval-augmented generation, that is a fair trade.
- Plan rungs used below. ScrapingBee: Freelance $49 (250,000 credits), Startup $99 (1,000,000), Business $249 (3,000,000). ZenRows: Build $39 (120,000), Launch $99 (500,000), Growth $399 (3,000,000). ScraperAPI: Startup $149 (1,000,000), Business $299 (3,000,000). Scrapfly: Pro $100 (1,000,000), Enterprise $500 (5,500,000). Firecrawl: Standard $99 (100,000).
- ZenRows says failed requests are not charged and free credits refresh monthly without rollover. Scrapfly has a
cost=trueflag that returns a request's cost without running it, handy for pricing your own targets. - Zyte tiers sites by difficulty. With a $100 monthly commitment, tier 1 drops to $0.10 (HTTP) and $0.75 (browser) per 1,000 (Zyte pricing).
- Bright Data and Oxylabs are for sites that beat the cheaper APIs. Bright Data's Scale plan is $499/mo with 383K requests included. Our read: you probably do not need either on day one.
- Different shapes. Apify sells ready-made scrapers (Actors). Browserbase rents whole browsers for AI agents and multi-step flows, not bulk crawling. Jina Reader returns markdown when you prefix a URL with
r.jina.ai/, at 20 requests per minute without a key.
Open source. Crawl4AI is Apache 2.0 and built for LLM pipelines; its README asks for an attribution such as a "Powered by Crawl4AI" badge (Crawl4AI). Scrapy is BSD-3-Clause (Scrapy). Crawlee is Apache 2.0, written in TypeScript, and drives Playwright or Puppeteer or plain HTTP with Cheerio (Crawlee). Our read: these are the right default for friendly sites and one-off crawls. What you do not get is proxies and unblocking, which is where your weekends go if targets fight back.
Decision table by job#
This is our editorial advice, based on the published prices and features above.
| Job | Our pick | Why |
|---|---|---|
| Feed docs or articles into RAG | Firecrawl, or Crawl4AI if self-hosting | Clean markdown out, fewer cleanup steps |
| Quick prototype, a few hundred URLs | Jina Reader or Firecrawl free | No plan, no card |
| Seed a directory website from known sites | Scrapy or Crawlee on a VPS | Friendly targets, lowest cost per page |
| Price monitoring on e-commerce sites | ZenRows, ScrapingBee or Scrapfly | Proxies and JS when needed, cheap base credits |
| Mostly static pages at volume, low budget | Zyte API pay as you go, or Scrapfly | $0.13 per 1,000 HTTP responses; 200,000 credits for $30 |
| Sites that block the cheaper APIs | Bright Data Web Unlocker or Oxylabs | Pay per success on hard targets |
| An existing scraper for a known platform | Apify Actors | Someone already maintains it |
| AI agent that logs in and clicks | Browserbase | Full browser sessions, billed by the hour |
| Lead lists of people | Think twice | Personal data brings in the GDPR (see below) |
Worked cost example: 100,000 pages a month#
Example, not a quote. Our arithmetic from the published monthly prices above, picking the smallest plan that covers the job. It ignores taxes, retries, overage and per-domain surcharges.
Scenario A: 100,000 static pages (1 credit each).
| Tool | Credits needed | Smallest plan that covers it | Monthly cost |
|---|---|---|---|
| Zyte API (tier 1, PAYG) | n/a | 100 x $0.13 | $13 |
| Scrapfly | 100,000 | Discovery (200,000) | $30 |
| ZenRows | 100,000 | Build (120,000) | $39 |
| ScrapingBee | 100,000 | Freelance (250,000) | $49 |
| ScraperAPI | 100,000 | Hobby (100,000) | $49 |
| Firecrawl | 100,000 | Standard (100,000) | $99 |
| Bright Data (PAYG) | n/a | 100 x $1.5 | $150 |
Scenario B: the same 100,000 pages, all JS-rendered.
| Tool | Credits needed | Smallest plan that covers it | Monthly cost |
|---|---|---|---|
| Firecrawl | 100,000 (no separate JS charge listed) | Standard (100,000) | $99 |
| ZenRows | 500,000 (5x) | Launch (500,000) | $99 |
| ScrapingBee | 500,000 (5 each) | Startup (1,000,000) | $99 |
| Scrapfly | 600,000 (6 each) | Pro (1,000,000) | $100 |
| Zyte API (tier 1, PAYG) | n/a | 100 x $1.01 | $101 |
| ScraperAPI | 1,000,000 (10 each) | Startup (1,000,000) | $149 |
| Bright Data (PAYG) | n/a (no separate JS charge listed) | 100 x $1.5 | $150 |
Scenario C: JS plus premium or residential proxies (25 to 30 credits each). ScrapingBee $249, ScraperAPI $299, ZenRows $399, Scrapfly $500, each on its 3M-plus credit plan.
Our read: the ranking flips. Firecrawl goes from the priciest credit API on static pages to level with the pack on rendered ones, and Zyte drops from cheapest to mid-table. Sample 50 real target URLs, see which need JS or premium proxies, and price that mix. If the text then goes into an LLM, the model bill may dwarf the scraping bill, so price both.
Legality and ethics#
Not legal advice. This section summarises primary and reputable sources so you know what to ask a lawyer. The law differs by country and turns on facts.
robots.txt is a signal, not a lock. The Robots Exclusion Protocol is now RFC 9309 (September 2022), and it says plainly: "These rules are not a form of access authorization" (RFC 9309). Our read: respect it anyway; it is the clearest statement of what the owner wants. For the other side, see how to block AI crawlers.
hiQ Labs v. LinkedIn (US Ninth Circuit, 2022). In an opinion filed 18 April 2022, on remand from the Supreme Court, the Ninth Circuit again affirmed a preliminary injunction stopping LinkedIn from denying hiQ access to public member profiles. The panel found hiQ had raised a serious question about whether the CFAA's "without authorization" concept applies where prior authorization is not generally required but a particular person or bot is refused access (Ninth Circuit opinion, No. 17-16783). It was a preliminary injunction ruling on the US anti-hacking statute, not a final ruling that scraping is legal.
Meta v. Bright Data (N.D. Cal., January 2024). Judge Edward Chen granted summary judgment to Bright Data on Meta's breach-of-contract claims, finding that its logged-out scraping of public data did not breach Meta's terms, which apply to "users" (Technology & Marketing Law Blog). Our read: the logged-out detail matters. Scraping behind a login you agreed to is different.
Personal data. If pages hold names, emails or profiles of people in the EU or UK, the GDPR applies to you as the collector, whatever the US cases say. Our GDPR checklist for solo SaaS founders covers the basics.
Our ethics checklist:
- Read robots.txt and the terms first, and do not log in to scrape without permission.
- Rate-limit yourself, and prefer an official API or export when one exists.
- Store only what you need, and drop personal data you did not mean to collect.
- Do not republish copyrighted content wholesale; facts, summaries and links are safer than copies.
Frequently asked questions#
What is the cheapest web scraping API for an indie developer?#
For plain HTML, Zyte API at $0.13 per 1,000 successful HTTP responses (simplest tier) or Scrapfly's 200,000 credits for $30/mo. For a $0 prototype, ZenRows gives 5,000 free credits a month. Once you need JS rendering or premium proxies, compare after the credit multipliers.
Which scraping API is best for RAG and AI apps?#
Firecrawl and Jina Reader both return clean markdown that is ready to chunk and embed. Firecrawl's free plan has 1,000 credits a month and a base scrape costs 1 credit per page. Jina Reader works by putting r.jina.ai in front of a URL and gives each new API key 10M free tokens. If you would rather self-host, Crawl4AI is open source under Apache 2.0.
How much does JavaScript rendering cost on scraping APIs?#
ScrapingBee and ZenRows charge 5 credits per JS-rendered request, Scrapfly adds 5 credits and ScraperAPI charges 10. Zyte charges $1.01 per 1,000 browser-rendered responses against $0.13 for HTTP at its simplest tier. Firecrawl and Bright Data list no separate JS charge.
Is web scraping legal?#
It depends on what you scrape, how, where and why. RFC 9309 says robots.txt rules are not a form of access authorization. In hiQ v. LinkedIn, the Ninth Circuit upheld a preliminary injunction over public profile data, and in Meta v. Bright Data a district court found logged-out scraping of public data did not breach Meta's terms. Personal data brings in the GDPR. This is not legal advice.
Should I use an open source scraper instead of an API?#
Use Scrapy (BSD-3-Clause), Crawlee or Crawl4AI (both Apache 2.0) if your targets are mostly friendly sites and you already run a server. Pay for an API when sites block you or you need residential proxies or real browsers at scale, because that is the hard part the vendors sell.
Launch what you build#
When your scraper-powered app or directory is ready, run through our product launch checklist for bootstrapped makers, then submit your product for free. For more tool guides like this one, join the ZeroRaised newsletter.
Sources#
- Firecrawl pricing, last checked 13 Nov 2026
- Firecrawl billing docs, last checked 13 Nov 2026
- ScrapingBee pricing, last checked 13 Nov 2026
- ScraperAPI pricing, last checked 13 Nov 2026
- ScraperAPI credits and requests docs, last checked 13 Nov 2026
- ZenRows pricing, last checked 13 Nov 2026
- Scrapfly pricing, last checked 13 Nov 2026
- Scrapfly billing docs, last checked 13 Nov 2026
- Zyte pricing, last checked 13 Nov 2026
- Bright Data Web Unlocker pricing, last checked 13 Nov 2026
- Oxylabs Web Scraper API pricing, last checked 13 Nov 2026
- Apify pricing, last checked 13 Nov 2026
- Browserbase pricing, last checked 13 Nov 2026
- Jina Reader, last checked 13 Nov 2026
- Crawl4AI on GitHub, last checked 13 Nov 2026
- Scrapy on GitHub, last checked 13 Nov 2026
- Crawlee on GitHub, last checked 13 Nov 2026
- RFC 9309: Robots Exclusion Protocol, last checked 13 Nov 2026
- hiQ Labs v. LinkedIn, Ninth Circuit opinion No. 17-16783 (filed 18 April 2022), last checked 13 Nov 2026
- Technology & Marketing Law Blog on Meta v. Bright Data, last checked 13 Nov 2026
Frequently asked questions
What is the cheapest web scraping API for an indie developer?
For plain HTML, Zyte API at $0.13 per 1,000 successful HTTP responses (simplest tier) or Scrapfly's 200,000 credits for $30/mo. For a $0 prototype, ZenRows gives 5,000 free credits a month. Once you need JS rendering or premium proxies, compare after the credit multipliers.
Which scraping API is best for RAG and AI apps?
Firecrawl and Jina Reader both return clean markdown that is ready to chunk and embed. Firecrawl's free plan has 1,000 credits a month and a base scrape costs 1 credit per page. Jina Reader works by putting r.jina.ai in front of a URL and gives each new API key 10M free tokens. If you would rather self-host, Crawl4AI is open source under Apache 2.0.
How much does JavaScript rendering cost on scraping APIs?
ScrapingBee and ZenRows charge 5 credits per JS-rendered request, Scrapfly adds 5 credits and ScraperAPI charges 10. Zyte charges $1.01 per 1,000 browser-rendered responses against $0.13 for HTTP at its simplest tier. Firecrawl and Bright Data list no separate JS charge.
Is web scraping legal?
It depends on what you scrape, how, where and why. RFC 9309 says robots.txt rules are not a form of access authorization. In hiQ v. LinkedIn, the Ninth Circuit upheld a preliminary injunction over public profile data, and in Meta v. Bright Data a district court found logged-out scraping of public data did not breach Meta's terms. Personal data brings in the GDPR. This is not legal advice.
Should I use an open source scraper instead of an API?
Use Scrapy (BSD-3-Clause), Crawlee or Crawl4AI (both Apache 2.0) if your targets are mostly friendly sites and you already run a server. Pay for an API when sites block you or you need residential proxies or real browsers at scale, because that is the hard part the vendors sell.