The Keepa API is a separate subscription from the Keepa Pro plan you use in your browser, and it is priced by token rate, not by seats or ASIN counts. Keepa’s plan picker lists four presets — Starter at 20 tokens per minute, Developer at 250, Business at 2,000 and Enterprise at 10,000, with a slider for rates in between (keepa.com plan picker, checked 2026-08-25). Third-party listings put Starter at €49 per month; Keepa’s own checkout is the only place to confirm that figure (see the pricing section below). One product costs one token, and a Starter plan mints 892,800 tokens over a 31-day month — roughly 892,000 plain product pulls, or about 47,000 pulls if you also request full marketplace offers.
This page covers the API only. For the browser extension, the charts and the Pro subscription, see the Keepa tool page .
Keepa API pricing: four presets and a slider
Keepa sells API access by the number of tokens your plan generates every minute. The tier names and token rates below are read off Keepa’s own plan picker (checked 2026-08-25). The euro prices are third-party: keepa.com puts its checkout behind a Cloudflare challenge that automated checks cannot pass, so the figures come from listings maintained by RevenueGeeks (updated 2026-07-22) and Front Desk Review (verified 2026-06-18), which agree on the entry price of €49. Treat them as indicative and read the number off Keepa’s checkout before you budget (data checked 2026-08-25).
| Tier | Token rate (official) | Monthly price (third-party) | Tokens per 31-day month | Burst capacity |
|---|---|---|---|---|
| Starter | 20 / minute | €49 | 892,800 | 1,200 |
| Developer | 250 / minute | €459 | 11,160,000 | 15,000 |
| Business | 2,000 / minute | €2,499 | 89,280,000 | 120,000 |
| Enterprise | 10,000 / minute | €11,099 | 446,400,000 | 600,000 |
The monthly and burst columns are arithmetic on the official token model, not separate claims: monthly tokens are rate × 60 × 24 × 31, and burst capacity is rate × 60.
Run those numbers against the prices and the effective rate drops sharply as you climb: about €5.49 per 100,000 tokens on Starter, €4.11 on Developer, €2.80 on Business and €2.49 on Enterprise. If your workload sits just above a tier’s ceiling, the next tier up is usually cheaper per token than adding a second small subscription.
Billing rules are documented rather than guessed (Keepa’s Plans & Tokens page, checked 2026-08-25):
- Plans are prepaid for one month with automatic renewal, paid by credit card through Stripe.
- Cancelling schedules termination at the end of the period you already paid for, and you can resume until it expires. After that, you have to subscribe again.
- Upgrades take effect immediately with a prorated charge. Downgrades are allowed only once every 28 days and return account credit, not a refund.
- Multiple subscriptions can run at once. They share one API key and their token rates add up, but each carries its own billing cycle.
- The page states plainly: “Keepa does not offer discounts.”
The API has no free entry point: every route in, the MCP server included, requires an active paid subscription (Keepa MCP documentation, checked 2026-08-25). The free browser extension and the paid Pro subscription are different products, and neither one is a trial of the API.
How Keepa API tokens actually work
Keepa meters the API with a token bucket rather than a monthly quota, which changes how you should plan a workload.
Your plan drips a fixed number of tokens into a bucket every minute, 24 hours a day, whether or not you make a single request. Every token is valid for 60 minutes, so the bucket never holds more than rate × 60 tokens — 1,200 on Starter, 3,600 on the 60-per-minute plan Keepa uses as its documentation example. Anything unused past the hour is gone: “Unused tokens expire after 60 minutes” (Keepa Plans & Tokens, checked 2026-08-25).
Two practical consequences follow, and both catch new subscribers out.
First, you cannot bank a month of tokens for one big weekend crawl. A Starter plan can spend at most 1,200 tokens in a single burst, then it refills at 20 per minute. A 50,000-ASIN catalogue refresh on Starter is not blocked, but it takes roughly 42 hours of steady drip, not an afternoon.
Second, an idle plan is a wasted plan. Tokens generated at 3 a.m. while your script sleeps expire at 4 a.m. Sellers who run one weekly sourcing pass pay for a full month of generation and use a fraction of it — that is the strongest argument for either scheduling work continuously or staying on the browser tools.
Every API response reports the bucket state, so you never have to guess: tokensLeft is your current balance, refillRate is your plan’s per-minute rate, refillIn is the milliseconds until the next refill, and tokensConsumed is what the call just cost. Keepa’s documentation notes that tokensLeft can go negative, because a request that starts with a positive balance is always executed. Run dry and the API answers HTTP 429 with the message “You are out of tokens” (Keepa request documentation, checked 2026-08-25).
What each request costs in tokens
One token per product is the headline, but the cost table is where budgets are actually won or lost. These figures come from Keepa’s API documentation index and the product endpoint page , checked 2026-08-25.
| Request | Token cost |
|---|---|
Product request (/product) | 1 per ASIN |
| Marketplace offers on a product | +6 per found offer page (max 10 offers per page) |
| Buy Box data on a product | +2 per product |
| Stock data in offers | +2 per product |
| Rating and review-count history | up to +1 per product |
| Historical variations | +1 per product with a parent ASIN |
Product search (/search?type=product) | 10 per result page (up to 10 results) |
Browsing deals (/deal) | 5 per 150 deals |
Best sellers (/bestsellers) | 50 per list (up to 100,000 ASINs) |
Most rated sellers (/topseller) | 50 per list |
Seller information (/seller) | 1 per seller |
| Category lookup or search | 1 per request or search |
Seller Finder (/sellerquery) | 10 + 1 per 100 returned seller IDs |
The stats parameter — current prices, minimum and maximum prices, weighted means — costs nothing extra. That is the single best value in the table: a plain product call plus statistics is still one token.
Offers are the opposite. A listing with 30 live offers spans three offer pages, so it costs 1 + 18 = 19 tokens. Pulling full offer data on 10,000 ASINs is not a 10,000-token job: assume two offer pages per listing and it is 10,000 x (1 + 12) = 130,000 tokens. Buy Box and stock flags add two tokens each on top.
Two parameters exist purely to save tokens. update=-1 tells the API not to refresh at all: if the ASIN is missing from Keepa’s database the request consumes zero tokens and returns nothing, which is how you screen a supplier list cheaply. Going the other way, update=0 forces live data and costs one extra token when the last refresh was under an hour old. Left alone, Keepa refreshes anything older than roughly an hour automatically: “If our last update is older than ~1 hour, it will be automatically refreshed before being delivered.”
Keepa also ships changes on a live API changelog
— the Seller Finder endpoint arrived on 2026-08-09 and an image variant field on 2026-08-18 — so pin your integration against that page rather than against a blog post’s snapshot.
The cost most people miss: tracking lowers your refill rate
Keepa’s Tracking API is where budgets quietly break. Trackings do not bill per notification; they permanently reduce the rate at which your bucket refills. Keepa’s tracking documentation states it directly: “Each tracked product decreases your token refill rate.”
The arithmetic, checked 2026-08-25:
- A regular tracking costs 0.9 tokens per update per locale tracked.
- A marketplace tracking (offers-level) costs 9 tokens per update per locale tracked.
- Adding a tracking costs 1 token; every other tracking operation costs zero.
- The reduction is rounded to an integer and refreshed every five minutes, reported as
tokenFlowReductionin each response.
Keepa’s own worked examples: 2,000 regular trackings on one locale at hourly updates consume 1,800 tokens per hour, which is 30 tokens per minute off your refill rate. And 700 marketplace trackings on one locale every 12 hours cost 8.75 tokens per minute, rounded to 9.
That first example is larger than a Starter plan’s entire output. As the documentation puts it, “You can track as many products as your token rate allows” — once the rate hits zero, no tracking and no token-consuming request goes through. Trackings also survive changes badly: downgrade to a plan that cannot fund the list and the excess trackings are deactivated, then removed after seven days unless reactivated.
Keepa’s own suggestion for anyone running both a monitoring list and a research pipeline is to keep two accounts with separate API subscriptions, so a growing tracking list cannot starve your ad-hoc queries.
Pulling Keepa data without writing code
“Do I need a developer?” is the real question behind most Keepa API searches. As of 2026-08-25 there are three honest answers, and only one of them is new.
Route one is Keepa’s hosted MCP server, documented at keepa.com/api-docs/mcp.html
(checked 2026-08-25). It connects an AI assistant — Claude Code, Claude Desktop, Cursor, VS Code, Gemini CLI, Windsurf, Codex CLI, Cline and other Model Context Protocol clients — straight to your Keepa data over a single endpoint, https://keepa.com/mcp, authenticated with your API key as a bearer token. It is fully hosted: “there is nothing to download, install or update.” Once connected, the assistant can look up products with statistics and Buy Box data, walk price histories, render Keepa charts, run Product Finder queries, browse deals and Lightning Deals, pull best-seller lists, inspect sellers and manage trackings — by asking in plain language. Three caveats belong next to that: it still requires an active paid API subscription, every tool call spends the same tokens as a direct request, and clients that authenticate remote servers only through OAuth (claude.ai custom connectors, and Claude Desktop’s connector UI) cannot connect yet. Keepa also warns that anyone holding the key can spend your tokens, so a checked-in config file is a live billing risk.
Route two is the Graph Image API, which returns a rendered price-history chart as an image from a URL. Anything that can display an image — a dashboard, a wiki, a spreadsheet cell — can show Keepa history without a single line of parsing code.
Route three is not the API at all. Keepa Pro’s in-browser Product Finder, Deals table and bulk Product Viewer already answer most sourcing questions and export to CSV, and they are covered on the Keepa tool page alongside the free extension. If your end state is a spreadsheet you look at once a week, the browser tools get you there without a token budget.
One route that is not on this list: third-party connectors that promise Keepa data inside a no-code builder still resell API access, so you are paying for tokens either way — plus their margin.
Do you need the API, or is Keepa Pro enough?
Start with what Pro already gives you. Keepa’s own account interface tells Pro subscribers: “Your Keepa Pro plan already generates 1 token per minute. Dedicated API plans start at 20 tokens per minute — 20× the rate — and are billed separately from Pro” (keepa.com, checked 2026-08-25). One token per minute is 1,440 products a day, or 44,640 in a 31-day month, at no cost beyond the Pro subscription itself. Third-party listings quote Pro between €19 and €29 per month depending on when they were checked, which is the same disagreement flagged on our Keepa tool page — another figure to confirm at checkout.
That single token per minute is the honest dividing line. Sizing by workflow:
- Weekly sourcing on a few hundred ASINs. Pro’s Product Finder and Deals table cover it. The API adds cost and no capability.
- Nightly catalogue refresh, no offers data. Pro’s rate tops out near 1,440 products a day. Starter generates 28,800 a day, which covers a catalogue of roughly 25,000 ASINs with headroom for a failed run; past that, size up.
- Repricing or Buy Box monitoring with offers data. Offer pages at six tokens each dominate the bill; size the plan on offers, not ASINs. Most repricing workloads land on Developer or above.
- Any live monitoring list. Budget the refill-rate reduction first and whatever is left over for queries second.
- One-off market research. Neither. A month of Pro, or a competing dataset, beats a month of API you will use for two days.
If you are weighing the whole toolset rather than the API alone, the price-history market has real alternatives at different price points — Keepa alternatives compares them for sellers, and CamelCamelCamel alternatives covers the free end. For reading the charts themselves, our older Keepa and CamelCamelCamel walkthrough still holds up.
Frequently asked questions
How much does the Keepa API cost?
Keepa’s plan picker lists Starter at 20 tokens per minute, Developer at 250, Business at 2,000 and Enterprise at 10,000, plus a slider for custom rates (checked 2026-08-25). Third-party listings put those tiers at €49, €459, €2,499 and €11,099 per month; keepa.com blocks automated price checks, so confirm at its checkout (data checked 2026-08-25).
Is there a free Keepa API tier?
No. The free browser extension and the paid Pro subscription do not include a dedicated API plan, although Keepa’s account interface states that a Pro plan generates 1 token per minute (checked 2026-08-25).
How many products can one token buy?
One token returns one product with price history, and adding the stats field costs nothing extra. Marketplace offers add six tokens per offer page of up to 10 offers, Buy Box and stock data add two tokens each.
Can I use the Keepa API without writing code?
Yes, through Keepa’s hosted MCP server, which connects AI assistants such as Claude, Cursor and Gemini CLI to your Keepa data with your API key. It still requires a paid API subscription and spends the same tokens per call (Keepa MCP documentation, checked 2026-08-25).
What happens if I run out of tokens?
The API returns HTTP 429 with “You are out of tokens” and your balance can show a negative number, since any request that starts with a positive balance is executed in full. The bucket refills at your plan’s rate every minute.