DEV Community: AgentStack Notes

Firecrawl Credits Explained: What 1,000 Free Credits a Month Buys

AgentStack Notes — Mon, 09 Nov 2026 16:22:30 +0000

Originally published at AgentStack Notes. Last fact-checked: November 9, 2026.

Disclosure: This article was drafted with AI assistance and checked against Firecrawl's official pages. We may receive a commission if you sign up through the Firecrawl † link marked † in this article, at no extra cost to you. See our Advertising Disclosure.

Firecrawl turns web pages into data for AI agents, and it bills in credits. The price list shows plan sizes, but what matters is how many credits one request burns. A plain page scrape is cheap; adding structured extraction or a search changes the math quickly.

Note on experience: we have not run production workloads on Firecrawl and make no claim about result quality, speed or reliability. This is a reading of the official pricing page and billing docs on November 9, 2026. Prices and credit costs can change.

1. Credits per operation, as documented

† marks a referral link (see the disclosure above). "Not stated" means the page we read did not say.

Operation	Firecrawl † credit cost
Scrape	1 credit per page
Crawl	1 credit per page
Map	1 credit per call
Search	2 credits per 10 results (rounded up per 10; scraped results add per-page costs)
JSON format (LLM extraction)	+4 credits per page, on top of the base cost
PDF parsing	+1 credit per PDF page
Enhanced proxy	Billed at the same 1 credit; an escalated retry is not charged separately
Interact (browser sessions)	2 credits per browser minute
Other add-ons	Prompt-injection check +4, zero data retention +1, PII redaction +4
Extract endpoint	Not stated as a separate credit line
Screenshots	Not stated

The billing docs say modifiers stack: JSON format plus zero data retention is 1 + 4 + 1 = 6 credits per page. A request that returns a document, including a 403 or 404 response, costs 1 credit plus option costs; a request that returns no document costs 0 credits, with a few listed exceptions such as the threat-protection scan.

2. Plans, as listed on the pricing page

Plan	Price per month	Credits per month	Concurrent browsers
Free	$0	1,000	2
Hobby	$19 ($16/mo billed yearly)	5,000	5
Standard	$99 ($83/mo billed yearly)	100,000	25
Growth	$399 ($333/mo billed yearly)	500,000	50
Scale	$749 ($599/mo billed yearly)	1,000,000	100

Free-plan rate limits are 10 requests per minute on scrape, map and search, and 2 per minute on crawl, agent and interact. The pricing page describes Free credits as recurring monthly, not a one-time grant, and says searches and scrapes share one balance. Plan credits do not roll over by default; Scale plans roll unused credits over for one month, and pay-as-you-go credits stay on the account until used but expire if the subscription is canceled.

3. What the credits buy (our arithmetic)

The figures below are our own calculations from the published numbers above, using the monthly price. They are not Firecrawl's claims, and they ignore retries and failures.

Workload	Credits each	Free plan (1,000 credits)	Cost per 1,000 units on Standard ($99 / 100,000 credits)
Plain page scrape	1	1,000 pages	$0.99
Scrape with JSON format	1 + 4 = 5	200 pages	$4.95
Search, 10 results, no scraping	2	500 searches	$1.98
Scrape a PDF page	1 + 1 = 2	500 PDF pages	$1.98

On Hobby ($19 for 5,000 credits) a plain scrape works out to $3.80 per 1,000 pages, and the same JSON-format scrape works out to $19.00 per 1,000 pages. Moving from Hobby to Standard cuts the per-credit price from $0.0038 to $0.00099, so heavy users pay far less per page than hobby users, while light users get no benefit from the bigger plan.

4. How to use this

Prototyping an agent: the Free plan's 1,000 credits cover roughly 1,000 plain pages or 200 JSON-format pages a month at the stated costs. That is enough to test a workflow, not to run one.
Watch the add-ons: JSON extraction multiplies the cost of a page by 5. If you only need clean text or markdown, skip it.
Searches add up: a search that also scrapes its results pays the search cost plus each scrape.
Plan for the reset: unused credits do not roll over, so buying a larger plan "just in case" wastes credits in slow months.

If you are comparing vendors rather than just Firecrawl, see ScraperAPI vs ScrapingBee, Free Tiers of Web-Data Tools Compared and Web-Data MCP Servers Compared. If you want to try Firecrawl's free plan, the Firecrawl † signup link supports this site.

5. What we could not confirm

Credit cost of the Extract endpoint, screenshots and stealth proxy: not stated on the pages we read.
Rate limits for paid plans (the billing doc points to a separate Rate Limits page we did not read).
The affiliate-program terms shown to us differ between Firecrawl's own pages and our partner dashboard; this article does not rely on them.
Success rates, speed and data quality. We have not measured them.
These pages were read through a web-fetch tool that summarises pages, not a manual read of the rendered pages; check the live pricing page before you buy.

Sources (checked November 9, 2026)

ScraperAPI vs ScrapingBee: What One Credit Actually Buys

AgentStack Notes — Sun, 08 Nov 2026 15:46:07 +0000

Originally published at AgentStack Notes. Last fact-checked: November 8, 2026.

Advertising disclosure: This article was drafted with AI assistance and checked against the vendors' official pages. We may receive a commission if you sign up through the ScraperAPI † link marked † in this article, at no extra cost to you. Our application to ScrapingBee's affiliate program is pending, so links to ScrapingBee go to the official site. We have not run either service ourselves; every number below comes from the vendors' own pricing and documentation pages. See our Advertising Disclosure.

ScraperAPI and ScrapingBee both sell a scraping API that you pay for in "credits". The credit is the unit on the price list, but the two vendors charge different numbers of credits for the same kind of request, so comparing plan sizes alone is misleading. This page lines up the published credit costs, the plan prices and the free trials, then does the arithmetic for a few request types.

Note on experience: we have not measured success rates, speed or result quality for either service. This is a pricing comparison only. Prices and credit costs were read on November 8, 2026 and can change.

1. Credits per request, as documented

† marks a referral link (see the disclosure above). "Not stated" means the page we read did not say.

Request type	ScraperAPI †	ScrapingBee
Plain request, no JavaScript rendering	1 credit	Pricing page: requests "start at 1 credit"; the API docs section we read did not give a figure for `render_js=false`
JavaScript rendering	Listed as +10 extra per request (no combined total printed on the page)	5 credits; rendering is on by default (`render_js` defaults to true)
Premium proxy, no rendering	+10 extra (listed as extra cost)	10 credits
Premium proxy + rendering	25 credits	25 credits
Hardest-site option	`ultra_premium`: 30 credits, or 75 with rendering; paid plans only	`stealth_proxy`: 75 credits per successful call
Marketplace sites (Amazon, Walmart, eBay)	5 credits per request	Not stated in the sections we read
Search engines (Google, Bing)	25 credits per request	Not stated in the sections we read
Social media (LinkedIn)	30 credits per request	Not stated in the sections we read
Bot-protected sites (Cloudflare, Datadome, PerimeterX, Akamai)	+10 credits per scrape when bypassed	Not stated in the sections we read
Failed requests	Charged for successful requests (status 200 and 404) and requests cancelled before 70 seconds; other failures not stated as charged	Not stated in general; in Auto-Mode, a request that fails every configuration costs 0 credits

Two things stand out. First, the default differs: a ScrapingBee request is a 5-credit request unless you turn rendering off, while a ScraperAPI request is 1 credit unless you turn rendering on. Second, ScraperAPI documents a domain-based cost table (marketplaces, search engines, social media) and a bot-protection surcharge; for ScrapingBee we did not find equivalent entries in the parts of the docs we read, so do not assume they are the same.

2. Plans and prices

Entry-level plans	Price per month	Credits	Concurrency
ScraperAPI Hobby	$49 ($44.10 billed yearly)	100,000	20 threads
ScraperAPI Startup	$149 ($134.10 billed yearly)	1,000,000	50 threads
ScraperAPI Business	$299 ($269.10 billed yearly)	3,000,000	100 threads
ScrapingBee Hobby	$19	75,000	25 concurrent
ScrapingBee Freelance	$49	250,000	50 concurrent
ScrapingBee Startup	$99	1,000,000	100 concurrent
ScrapingBee Business	$249	3,000,000	200 concurrent

Notes from the pages: ScrapingBee prices exclude VAT, and its credits are valid for one month and do not roll over. ScraperAPI's page lists larger plans (up to Advanced at $1,975 per month) and says pay-as-you-go is available from the Scaling plan; the page we read did not state a credit-rollover rule or whether prices include tax. The ScraperAPI page also says geotargeting on the two smallest plans is limited to the US and EU.

3. Price per 1,000 requests, from the published numbers

The figures below are our arithmetic, not vendor claims. We take the plan price, divide by the credit allowance, then multiply by the credit cost of the request type. They compare the roughly $49 plans (ScraperAPI Hobby, ScrapingBee Freelance) and ignore any tax, overage or discount.

Request type	ScraperAPI Hobby ($49 / 100,000 credits)	ScrapingBee Freelance ($49 / 250,000 credits)
Cost of one credit	$0.00049	$0.000196
ScraperAPI-style plain request (1 credit)	$0.49 per 1,000	Not comparable: ScrapingBee's default is 5 credits
ScrapingBee default request (5 credits)	Not the same product; ScraperAPI lists rendering as an extra cost	$0.98 per 1,000 (50,000 requests per month)
Premium proxy, no rendering (10 credits)	$4.90 per 1,000 (if the +10 is read as the full cost; see the note below)	$1.96 per 1,000 (25,000 requests per month)
Premium proxy + rendering (25 credits)	$12.25 per 1,000 (4,000 requests per month)	$4.90 per 1,000 (10,000 requests per month)

Note: ScraperAPI's page lists premium as "+10 extra" on top of the 1-credit base, so a premium request without rendering could be 10 or 11 credits. We show 10 and mark the reading; the 25-credit combination is stated explicitly. The same ambiguity applies to rendering alone, which is why we do not put a price on it.

On the one row where both vendors print the same total (premium proxy with rendering, 25 credits), the arithmetic above gives ScrapingBee Freelance a lower price per request at the roughly $49 tier. That is a pricing statement only. It says nothing about which service gets through a given site more reliably, and the cheaper credit may not stay cheaper if your target needs a more expensive option such as stealth or ultra-premium.

4. Free trials

ScraperAPI †: the pricing page advertises a 7-day trial with 5,000 API credits and no credit card. Its FAQ separately mentions a free plan of 1,000 credits with at most 5 concurrent connections; the page does not explain how the two relate.
ScrapingBee: 1,000 free API credits, no credit card required.

At the default costs above, 1,000 ScrapingBee credits is 200 default requests (5 credits each), while 1,000 ScraperAPI credits is 1,000 plain requests. For a trial, count requests, not credits.

5. A neutral way to choose

Your targets are mostly static pages: compare the plain-request rows. ScraperAPI's 1-credit plain request is the cheapest unit in this article, but ScrapingBee lets you turn rendering off too (its plain cost is not printed in the docs we read).
Your targets need JavaScript rendering or premium proxies: use the 25-credit row, where both vendors print the same total, and compare plan sizes from there.
You scrape marketplaces or search engines: ScraperAPI documents fixed credit costs for them (5 and 25). Check ScrapingBee's dedicated endpoints before you compare.
You want to test before paying: both have a no-card trial. Run your own target URLs and read the credit cost from the response; ScraperAPI documents an sa-credit-cost response header and a cost-check endpoint.

If you are choosing between platforms rather than scraping APIs, see Free Tiers of Web-Data Tools Compared, Web-Data MCP Servers Compared and Make vs Apify.

6. What we could not confirm

ScrapingBee's credit cost for render_js=false, whether failed requests are billed, and any Google-search endpoint cost: not found in the sections of the docs we read.
Whether ScraperAPI's render-only and premium-only costs are the extra amount alone or the extra amount plus the 1-credit base.
ScraperAPI credit rollover and tax treatment, and how its 5,000-credit trial relates to the 1,000-credit free plan.
Success rates, speed, and data quality of either service. We have not measured them.
These pages were read through a web-fetch tool that summarises pages, not a manual read of the rendered pages; check the live pricing pages before you buy.

Sources (checked November 8, 2026)

Web-Data MCP Servers for AI Agents, Compared

AgentStack Notes — Sun, 08 Nov 2026 10:26:12 +0000

Originally published at AgentStack Notes. Last fact-checked: November 8, 2026.

Disclosure: This article was drafted with AI assistance and checked against each vendor's official documentation. We may receive a commission if you sign up through the ScraperAPI † or Firecrawl † or Bright Data † links marked † below, at no extra cost to you. We have not tested these servers ourselves. Our applications to Apify and ScrapingBee are still pending, so links to those go to the official sites. See our Advertising Disclosure.

The Model Context Protocol (MCP) lets an AI client such as Claude Desktop, Cursor or VS Code call external tools. Most web-data vendors now ship an MCP server, so an agent can fetch a page or run a search without glue code. The servers differ in how you connect, how you sign in, and how much of the tool list is documented. This page compares five of them.

Note on experience: we have not run any of these servers in production. Facts below are from the vendor pages linked in the sources, read on November 8, 2026. Where a page did not say, we write "not stated" rather than guess.

1. At a glance

† marks a referral link.

Vendor	Remote endpoint (as documented)	Sign-in	Documented free allowance
Apify	`https://mcp.apify.com`	OAuth in the browser, or a bearer token; docs and Actor-search tools work without a token	Free plan: $5 of platform usage per month; the MCP page gives no separate MCP pricing
Firecrawl †	`https://mcp.firecrawl.dev/v2/mcp-oauth` (OAuth) or `https://mcp.firecrawl.dev/v2/mcp` (API key)	Keyless, browser sign-in, or API key	Keyless: Search, Scrape and Parse "within daily limits" (numbers not given). Free plan: 1,000 credits per month
Bright Data †	`https://mcp.brightdata.com/mcp?token=YOUR_API_TOKEN` (from the project README)	Bright Data API token	5,000 requests per month, no card to start, no rollover
ScraperAPI †	`https://mcp.scraperapi.com/mcp`, or self-hosted with Python or Docker	Not stated on the page we read	Not stated for MCP; the regular plan is a 7-day, 5,000-credit trial, then 1,000 credits per month
ScrapingBee	`https://mcp.scrapingbee.com/mcp`	API key as the `api_key` query parameter	Same credits as the underlying API endpoint, no MCP surcharge; 1,000 free credits on the regular trial

2. Tools each server documents

Apify: enabled by default are search-actors, fetch-actor-details, call-actor, apify/rag-web-browser, apify/web-fetch, search-apify-docs, fetch-apify-docs and report-problem. Optional categories cover Actor runs, datasets, key-value stores, tasks and schedules. The rate limit is stated as up to 30 requests per second per user (HTTP 429 beyond that).
Firecrawl: firecrawl_scrape, firecrawl_map, firecrawl_search, firecrawl_parse, firecrawl_crawl, firecrawl_check_crawl_status, firecrawl_agent, firecrawl_agent_status, firecrawl_interact and firecrawl_interact_stop, plus groups for research, developer search, government search and monitors. The old Extract tool is described as deprecated.
Bright Data: 30 tools in total according to the project README, in groups named scrapers (6), browser (14), geo (2) and advanced_scraping (3). Every group also includes search_engine, scrape_as_markdown and ask_brightdata_assistant. You can limit the exposed tools with groups or tools parameters to keep the client's context small.
ScraperAPI: the MCP page says it supports ScraperAPI parameters such as geotargeting, premium proxies and JavaScript rendering. It does not list tool names.
ScrapingBee: HTML tools (get_page_text, get_page_html, extract_page_data, get_screenshot, get_file), search and AI tools (fast_search, get_google_search_results, ask_chatgpt, ask_gemini), Amazon and Walmart tools, YouTube tools, and get_scrapingbee_usage (limited to 6 calls per minute).

3. Which clients are named

Apify: Claude Desktop and VS Code extensions with MCP support; any client implementing the MCP specification should work. The page also points to a plugin for coding agents such as Claude Code, Cursor, GitHub Copilot and Codex.
Firecrawl: ChatGPT and Codex, Claude (including Claude Desktop and Claude Code), GitHub Copilot CLI and VS Code.
Bright Data (README): Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, Gemini CLI, Zed and Warp, plus any client that supports local MCP servers.
ScrapingBee: Claude Desktop and Cursor via npx mcp-remote, ChatGPT through a custom app (plan limits apply), and any client that supports MCP over streamable HTTP.
ScraperAPI: not stated on the page we read.

4. Things to check before you connect

Where the key lives. ScrapingBee's page notes that the API key sits in the URL, so keep that config out of source control and rotate the key if it leaks. Bright Data's remote URL, as written in its README, also carries a token in the query string. Apify's default flow uses OAuth so the token stays out of the client config.
How many tools your client loads. A long tool list takes context. Bright Data documents group and tool filters for this reason.
Which meter runs. Apify bills platform usage, Firecrawl bills credits, Bright Data counts requests, ScrapingBee charges the same credits as the matching API call.
Concurrency and rate limits. ScrapingBee's hosted server has a shared global concurrency limit and may return HTTP 429; Apify states 30 requests per second per user.

5. A neutral way to choose

Try an MCP server without an account: Firecrawl documents a keyless mode with daily limits, and Apify's docs and Actor-search tools work without a token.
Most documented tool surface and filters: Bright Data lists 30 tools in named groups.
A specific site or marketplace: ScrapingBee lists Amazon, Walmart and YouTube tools; Apify's Actor catalog is its equivalent.
You already use an HTTP scraping API and want the same engine in your agent: ScraperAPI † and ScrapingBee both document MCP servers on top of their APIs.

6. What we could not confirm

ScraperAPI MCP: tool names, supported clients, authentication and credit usage.
Bright Data: the product page and the docs overview did not state the endpoint, tool count or authentication; those details come from the project README.
Firecrawl: credit cost per MCP tool (only Alexandria discovery is stated as free) and the numeric keyless daily limits.
Real-world reliability, latency and result quality of any of these servers; we have not measured them.

Sources (checked November 8, 2026)

Make vs Apify: Which Should You Try First to Add Web Data to Your AI Agent?

AgentStack Notes — Sun, 08 Nov 2026 02:37:14 +0000

Originally published at AgentStack Notes. Last fact-checked: November 7, 2026.

Disclosure: This article was drafted with AI assistance and checked against each platform's official pages. It may contain affiliate links; we have applied to the affiliate programs mentioned below and the applications are pending, so today every link goes to the official sites, not to tracked referral links. If that changes and we earn a commission, it costs you nothing extra. See our Advertising Disclosure.

If you're building an AI agent or an automated workflow and you've hit the point where it needs to pull data from the live web, you've probably run into two names over and over: Make (a no-code automation hub) and Apify (a web-scraping and browser-automation platform). They get mentioned in the same breath so often that it's easy to assume they're interchangeable. They aren't, and picking the wrong one first can cost you a week of rebuilding your workflow.

This guide compares them on the two things that actually decide which one you should try first: what each one is built to do, and what it costs you to find that out.

Note on experience: we have hands-on, daily experience running browser automation and web-scraping workloads (headless browsers, CDP-level control, anti-bot handling) as part of real automation projects. We have not personally run production workloads on Make or Apify specifically. Everything below about their features, pricing, and affiliate terms is sourced directly from their official documentation and pricing pages (linked throughout, checked November 7, 2026) — not from first-hand testing of these two platforms. Treat the general automation judgment as experience-based, and the platform-specific numbers as citation-based.

1. What each tool is actually for

Make is a general-purpose, no-code automation hub. You connect apps (Gmail, Notion, Slack, your CRM, a webhook from your AI agent, and so on) with a visual flowchart-style builder, and it runs the whole chain for you. Web scraping is one module you can drop into a larger flow — it is not the product's center of gravity.

Apify is built specifically around collecting data from the web and running browser automation at scale. It ships as a marketplace of pre-built "Actors" (ready-made scrapers for sites like Google Search, Instagram, Amazon, etc.) plus an SDK if you want to write your own. If your agent's core job is "go get this data from the web," Apify is closer to a purpose-built tool for that job.

A simple way to decide: if you need to orchestrate a multi-step workflow where web data is one ingredient among several (trigger → scrape → summarize with an LLM → post to Slack), Make's visual builder will likely get you there faster. If the web data collection itself is the hard part (JavaScript-heavy sites, anti-bot measures, pagination at scale, proxy rotation), Apify's purpose-built Actors and infrastructure are the more direct fit.

2. What it costs to actually try them

	Make	Apify
Free plan price	$0/month	$0/month
Free plan allowance	1,000 credits/month (official pricing page)	$5 of platform usage/month, metered at $0.20 per Compute Unit (official pricing page)
Free plan concurrency limit	Not stated on the pricing page we checked	5 concurrent runs (official pricing page)
Cheapest paid plan	Not confirmed on the page we checked — third-party sources cite roughly $9–$12/month, unverified against the official page	Starter: $19/month ($17/month billed annually) (official pricing page)

Both free plans are real free plans (not trials), but they are structured differently. Make counts every action in your flow as a "credit" — a flow with many steps eats through 1,000 credits faster than you'd expect. Apify's free plan is a $5/month usage credit metered per Compute Unit, so the real question is how much CPU/RAM-time your specific scraping job burns, which varies a lot by target site and Actor.

Practical takeaway: if you want to test a single, well-defined scraping task, Apify's $5 free usage is a more predictable way to find your actual cost per run. If you want to test an end-to-end workflow with several apps involved, Make's 1,000 free credits/month lets you build and run the whole chain before paying anything.

3. What you'd be signing up for as an affiliate (if you later want to recommend one)

Since this site covers both building with these tools and the economics of recommending them, here's what each program's official terms say, as of November 7, 2026:

Make Affiliate Program: 35% commission on referred subscription payments for 12 months from signup. Payout requires a minimum of $100 and 3 unique paying referred users in the same period; paid exclusively via Wise. No stated application review — any Make account holder can join.
Apify Affiliate Program: 20% of a referred customer's payments for their first 3 months, then 30% after that, capped at $2,500 per referred customer (higher cap for affiliates with 50+ active referrals). Monthly payout minimum is $100 (rolls over if unmet), plus a cumulative 3 paid referrals requirement. Paid via PayPal, bank transfer, or Wise, within 15 business days of month-end.

Neither program's official pages state a residency restriction, but neither confirms unrestricted eligibility either — if you're applying from outside the US/EU, confirm directly with each program before counting on approval.

4. Who each one suits

Try Make first if: your AI agent needs to talk to several other apps and services, web data is just one piece of a bigger chain, and you'd rather drag-and-drop a flow than write scraping logic.

Try Apify first if: the core problem is reliably getting structured data off of specific, possibly JavaScript-heavy or anti-bot-protected sites, and you're willing to work with (or write) purpose-built Actors to do it.

Try neither yet if: the site you need data from has a documented public API. Several experienced scrapers (see the Indie Hackers thread we reviewed while researching this) recommend checking for a usable backend API before reaching for any scraping tool — it's cheaper and far more stable than parsing HTML.

5. What we haven't verified yet

To keep this article honest about its own limits:

We have not independently load-tested either platform's reliability at scale.
Make's paid-tier pricing above the free plan is not confirmed from Make's own pricing page in this pass — treat the $9–$12/month figures as third-party estimates until verified directly.
Neither affiliate program's official pages confirm a residency restriction or cookie/attribution window; both are marked "unconfirmed" pending direct contact with each program.

If you've used either platform in production, what tipped your decision? We'll update this comparison as we get hands-on with both.