
ZooData
ZooData is an agent-native commerce intelligence API returning clean JSON on over 500 million Amazon and TikTok Shop products, with real-time pricing, two years of history and category-level market analysis for teams building AI agents.
What is ZooData?
ZooData describes itself as the data layer for AI agents: a commerce data service meant to be called by software rather than read by a person. Its output is structured JSON and its billing is per API call rather than per seat, on the argument that an agent works around the clock while a human analyst does not.
The acceptable use policy splits the offer in two. Data Collection covers agent-driven web retrieval; Domain Intelligence covers pre-processed vertical data, starting with e-commerce, with further verticals announced. The product argument is framed as the "4C": Clean, meaning flat JSON with an announced 80% cut in LLM parsing cost; Continuous, meaning real time plus more than two years of history; Comprehended, meaning sentiment, keywords and attributes extracted upstream for a claimed 95% token saving; and Commerce-native.
Six v2 endpoints are documented: product search, competitor lookup, market search, category taxonomy, real-time product and product history. Fashion Tags, which turns an image into structured attributes, is listed as coming soon. Freshness is published in tiers: Best Sellers Rank every 15 minutes, key prices every 30 minutes, priority data daily, the full catalog weekly. The headline figures are 500 million+ tracked products, roughly one second of response time and a 99.9% availability SLA. Commerce intelligence covers Amazon and TikTok Shop, while the public playground also accepts Walmart pages and news articles for generic collection.
Integration is deliberately standard rather than proprietary: an OpenAPI 3.0 specification imported in one click into LangChain, CrewAI, AutoGen, Claude MCP and OpenAI, plus an MCP server. Agent Skills install with a single npx command and ship six analysis modules and fourteen product-selection presets.
The home page names its rivals openly, positioning ZooData against data providers such as Oxylabs, ScrapingBee and Bright Data on one side, and against human-facing platforms such as Jungle Scout, Helium 10 and Keepa on the other. The About page states the thesis plainly: models are commoditizing, data is the new moat. The publisher is based in Sammamish, Washington, with a Discord community, its Skills code on GitHub and an X account.
What it does
- Search more than 500 million Amazon and TikTok Shop products with over 40 filters
- Analyze a category's market signals: demand, competitive density and margin benchmarks, refreshed daily
- Look up competitors by brand, ASIN, seller or keyword
- Pull real-time price, stock and Best Sellers Rank data, served without a cache
- Retrieve more than two years of history on BSR, price, sales and ratings
- Collect any web page through the scrape endpoint, in JSON, Markdown or raw HTML
- Call an OpenAI-compatible generative gateway billed per token
When to use ZooData / When not to
A quick filter to help you decide if ZooData is the right fit.
When to use ZooData
- Engineering teams building AI agents for product selection, competitive monitoring or dynamic pricing
- Amazon and TikTok Shop sellers and agencies that want to industrialize product research
- Developers who want ready-to-consume JSON instead of HTML they have to parse themselves
- High-volume operations: the site claims 10,000+ products a day against roughly 100 handled manually
- Teams already working with OpenAPI 3.0, MCP, LangChain, CrewAI or AutoGen
When not to use ZooData
- Anyone looking for a pricing SaaS or a human dashboard: the site states plainly "This is not a pricing SaaS - it's the data layer for your pricing Agent."
- Sellers working outside Amazon and TikTok Shop, the only two marketplaces the commerce intelligence layer covers today
- Businesses that need to redistribute or resell raw API responses, which the acceptable use policy forbids
- Teams that want to retrain AI or ML models on the retrieved data, prohibited without explicit written permission
- Non-technical users expecting a mobile app or a browser extension: there is none, and the full API reference sits behind an account
How to use ZooData
A typical end-to-end flow, from setup to results.
- Start with the public playground: the scrape endpoint can be tested on a live page, with JSON, Markdown or raw HTML output, before any account is created
- Create an account and collect the 1,000 bonus credits, with no credit card required
- Generate an API key from the console, in the hms_live_xxx format, and manage it from the API keys page
- Send a first POST request, for example to the product search endpoint, with an Authorization: Bearer header and Content-Type: application/json; the site advertises five minutes from zero to first call
- Import the OpenAPI 3.0 specification into LangChain, CrewAI, AutoGen or OpenAI to expose the endpoints as agent tools
- For Claude, add an mcpServers block pointing at the ZooData MCP server, carrying the same Authorization header
- Install the Agent Skills with npx skills add SerendipityOneInc/ZooData-Skills, or a narrower package for Amazon analysis or for the data layer alone
- Drive the skill from its CLI commands: categories, market, products, competitors, product, report, opportunity and check
- Use the web console for manual research when no agent is involved
- For generative workloads, ask sales to activate a USD balance, then call the OpenAI-compatible chat completions gateway with any compatible SDK
Pros & Cons
Pros
- Output designed for LLM consumption: flat JSON with insights already extracted, and a claimed reduction in token spend
- Rare historical depth at this price point, with more than two years of BSR, price, sales and rating history
- High freshness, documented tier by tier rather than merely claimed: BSR every 15 minutes, key prices every 30 minutes
- Pay-per-call pricing with no seats and no mandatory subscription, with volume discounts reaching 78% on the ultimate pack
- 1,000 credits given at sign-up, without a credit card, and those credits never expire
- Standard integration through OpenAPI 3.0 and MCP rather than a proprietary SDK, with Agent Skills installable in a single command
- A public playground allows testing before any account is created, and the documentation is unusually frank about what the service is not
Cons
- Complete silence on the GDPR: no mention, no DPA, no EU representative, no sub-processor list and no hosting country
- A very short privacy policy, eight sections, for a service whose business is collecting data
- The full API reference is unavailable without an account, since the api-docs page redirects to the login screen
- Commerce intelligence is limited to Amazon and TikTok Shop, with other verticals only announced
- Generative gateway pricing is not published: contacting sales is mandatory
- Pack credits expire six months after purchase, and credit and USD balances are never interchangeable
- A very young service with no exact legal name and no full postal address on the site: the domain was registered in June 2026 and first archived in July 2026
Pricing & Plans
A free entry point exists: a one-off bonus of 1,000 credits on registration, with no credit card required and no expiry, alongside a permanent Free plan on the web console at $0/month. Two separate balances coexist and are never interchangeable. Credits fund the data APIs: top-ups start at $10 for 5,000+ credits, that is $2.00 per 1,000 credits, and most endpoints cost 1 credit per call while market analysis costs 5. Prepaid packs run from $69 to $4,499, with the unit price falling from $1.38 to $0.45 per 1,000 credits. A separate USD balance funds the OpenAI-compatible generative gateway, billed per token and priced on quotation. Payment is by Visa, MasterCard or American Express through Stripe, as one-off purchases rather than a recurring subscription. The lowest paid entry point is therefore $10 in credits, or $29/month for the Pro plan of the web console.
- 1
- 000 bonus credits
- one-off
- no credit card
- 30 requests per minute
- 5 concurrent requests
- credits that never expire
- from $10
- 5
- 000+ credits
- $2.00 per 1
- 000 credits
- no expiry
- 120 requests per minute
- 10 concurrent requests
- $69
- 50
- 000 credits
- $1.38 per 1
- 000
- valid 6 months
- 120 requests per minute
- 10 concurrent requests
- $299
- 300
- 000 credits
- $1.00 per 1
- 000
- valid 6 months
- 180 requests per minute
- 20 concurrent requests
- $999
- 1
- 500
- 000 credits
- $0.67 per 1
- 000
- valid 6 months
- 300 requests per minute
- $4
- 499
- 10
- 000
- 000 credits
- $0.45 per 1
- 000
- valid 6 months
- $0/month
- basic search access
- limited exports
- $29/month
- unlimited search access
- unlimited exports
- priority support
- with annual billing announced at -20%
- billed per token in USD
- tiered by model and by volume
- priced on quotation once sales has activated the USD balance
Data, GDPR & hosting
A consolidated view of how ZooData handles your data.
GDPR overview
There is nothing to report, and that is itself the finding. Across the sixteen pages archived for this review, terms, privacy policy and acceptable use policy included, the words GDPR, CCPA, data processing agreement, sub-processor and opt-out do not appear once, neither in the visible text nor in the HTML. No Article 27 EU representative is named, no DPA is published or offered on request, and no formal list of sub-processors exists: only Firebase and Stripe are mentioned in passing. No hosting country or region is disclosed. The terms place the contract under the law of the State of Delaware, with thirty days of good-faith negotiation before any legal action. Rights of access, rectification and deletion are announced, but without reference to any regulatory framework. A European controller should treat this silence as an open question rather than as compliance.
Who owns the data?
ZooData's terms draw a clear line. The personal data you provide, namely email, name, usage and query records, payment details handled through Stripe and records of your consent to the acceptable use policy, remains yours: access, correction and deletion can be requested at any time from support@zoodata.ai. The data returned by the API is another matter. ZooData grants the customer no rights in third-party content, including product listings, images, trademarks and copyrighted text owned by the e-commerce platforms referenced through the service. Deciding whether a given use of that content is lawful is explicitly the customer's responsibility. No regulatory framework is invoked anywhere to qualify these rights.
Reuse rights
ZooData states that the data it collects serves to provide and improve the service, process transactions, send technical and support notifications, answer questions and keep compliance records. Two third parties are named: Firebase for authentication and Stripe for payments, each operating under its own policy. Cookies handle sessions and preferences and can be controlled from the browser. Nothing on the site says whether ZooData itself trains models on customer data. On the customer's side the acceptable use policy is restrictive: the retrieved data may not be used to train AI or ML models, to identify or profile individuals, or to reproduce copyrighted content beyond internal analysis, and raw responses may not be redistributed or resold. Reuse is therefore permitted for your own analysis, not for republication.
Data retention & training
Hosting summary
The site publishes nothing about where data is hosted. No country, no region, no data center and no cloud provider for the platform itself is named on any of the sixteen pages reviewed. The privacy policy limits itself to a single sentence on the subject, stating that appropriate security measures are implemented to protect personal information, without one technical or geographic detail. The only third parties identified anywhere are Firebase for authentication and Stripe for payments, each governed by its own policy, and neither is presented as part of a sub-processor list. What can be observed from outside is that the infrastructure is served behind Cloudflare, on an anycast address; that is a network observation made during this review, not a statement by ZooData. The only jurisdictional anchors the company provides itself are its United States base in Sammamish, Washington, and terms governed by the law of the State of Delaware. Anyone with a data residency requirement will have to ask directly, because the published material does not answer the question.
Things to keep in mind
Risks and trade-offs to weigh before adopting ZooData.
- No mention of the GDPR, no DPA and no EU representative: a blocking point for any European data controller who has to document its processing chain
- No hosting country or region is disclosed, so there is no way to know where the data physically sits before asking
- Legal responsibility for the use made of the collected data falls entirely on the customer, who must also respect the terms of the source platforms
- Contractual bans on retraining a model on the retrieved data and on redistributing raw responses can quietly conflict with a product decision taken later
- Pack credits are lost after six months and the credit and USD balances are compartmentalized: buying more volume than you will consume is a real risk
- Generative gateway pricing is not published, which makes budgeting impossible before a sales conversation
- Automating research at thousands of products a day makes it tempting to stop checking: an agent's ranking is only as good as the data behind it, and delegating category judgement entirely to a pipeline erodes the market intuition that used to catch obvious mistakes
Setup & Integrations
Technical difficulty
Low for a developer, out of reach for a non-technical buyer. The site advertises five minutes from sign-up to first call: create an account, generate a Bearer key, send a POST request. There is no no-code interface for the API, since every call is an HTTP JSON request, but the OpenAPI 3.0 specification imports into agent frameworks in one click, an MCP block can be copied straight into Claude, and the Agent Skills install with a single npx command. A public playground allows a first test without an account, and the web console covers manual research without code.
Deployment
Integrations
Supported languages
Behind ZooData
Fundraising
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement ZooData.
Frequently asked questions
Which marketplaces does ZooData's commerce intelligence cover?
Is there a free tier?
How much does one API call cost?
Do the credits expire?
How do I connect ZooData to an agent framework?
Can I test the service before creating an account?
Can I resell the data or train a model on it?
Does the site say anything about the GDPR?
How fresh is the data?
Which payment methods are accepted?
Should you pick ZooData?
ZooData is unusually clear about what it is. It sells a data layer for agents, not a dashboard for analysts, and every design decision follows from that: flat JSON with insights already extracted, an OpenAPI 3.0 specification and an MCP server instead of a proprietary SDK, and billing per API call rather than per seat, because an agent works around the clock.
On its own ground the offer holds up. Freshness is documented tier by tier rather than merely claimed, historical depth reaches beyond two years on rank, price, sales and ratings, and the volume discounts run deep enough, down to $0.45 per 1,000 credits, to make high-frequency agent workloads plausible. A public playground and 1,000 non-expiring free credits make the first evaluation genuinely risk-free.
The reservations are just as clear. The site says nothing at all about the GDPR: no mention, no data processing agreement, no EU representative, no sub-processor list, no hosting country. For a European controller feeding third-party marketplace data into its own systems, that silence is a question to settle before signing, not a detail. Commerce intelligence still covers only Amazon and TikTok Shop, the full API reference sits behind an account, generative gateway pricing requires a sales conversation, and the operation is very young: the domain was registered in June 2026, and no exact legal name or full postal address appears anywhere on the site.
The natural audience is technical: engineering teams wiring agents to live market data, and high-volume Amazon or TikTok Shop sellers who have outgrown manual research. For them the entry cost is close to zero and the evaluation is quick. For anyone carrying a compliance obligation, the questions come first.
- Choosing a selection results in a full page refresh.
- Opens in a new window.