Zyte review

Zyte is a set of hosted web-scraping services, not a single crawler library. Zyte API is a hosted HTTP API that takes a URL and extraction options, handles website access, and returns a response that can include raw HTML or structured fields. Zyte says the API selects an access strategy and escalates when a site demands it; the homepage says billing is on success.

Zyte Data is a managed collection and delivery service: Zyte’s team builds and maintains the collection and sends deduplicated, schema-ready data to a warehouse on a chosen cadence. Scrapy Cloud is hosted infrastructure for running Scrapy crawlers. Scrapy itself is the open-source framework for teams building their own crawlers. Zyte also lists an MCP Server for use from MCP clients. In a stack, you can call the API from an application or pipeline, hand collection to Zyte Data, or run your own crawler code on Scrapy Cloud.

The API uses HTTP POST requests to https://api.zyte.com/v1/extract. Zyte’s example authenticates with a Zyte API key using HTTP Basic authentication, then sends JSON containing a URL and requested outputs such as httpResponseBody and product. The response is JSON and can include status, extraction time, HTML and structured product fields.

The homepage lists Python, C#, cURL, Java, PHP, CLI, Node.js, Python Client, Scrapy and Ruby as integration or example options. It also links an API reference, getting-started guides, integration recipes and a change log. Zyte maintains Scrapy on GitHub, and the MCP Server is presented as an integration for MCP clients. Confirm the current SDK support and implementation details in Zyte’s docs before selecting a client library; language options shown on the homepage do not by themselves establish the scope of each SDK. The API and Scrapy are the clearest integration surfaces for application and crawler code.

Zyte lists no free API plan. The API offer includes a $5 free trial credit for 30 days. Scrapy is open source, but it is a framework for building crawlers, not a free allotment of Zyte API requests or managed data service.

Pay as you go has no commitment: HTTP requests cost $0.13–$1.27 per 1,000 and browser requests $1.01–$16.08, across five website pricing tiers. Monthly commitments start at $100, with HTTP at $0.10–$0.95 and browser at $0.75–$12.00 per 1,000. The $200 commitment lists $0.08–$0.76 HTTP and $0.60–$9.60 browser; the $500 commitment lists $0.06–$0.61 HTTP and $0.48–$7.68 browser. Enterprise pricing is custom.

The bill depends on request type, target-site tier and usage. Usage above a commitment is charged at the current discounted rate, with no overage penalty. Zyte says some advanced features may have additional usage-based costs, but does not publish those amounts.

Scrapy is an open-source framework that Zyte maintains, and the company links its GitHub repository. Zyte’s homepage does not name the license, so check the repository’s license file before adopting or distributing it. Scrapy lets a team own its crawler code; Scrapy Cloud is the hosted option for running those crawlers without owning servers.

Zyte API and Zyte Data are services rather than self-hosted software. The vendor describes data delivery to a warehouse, but does not describe a standard export or migration process for managed collections. A move away from the API means replacing its access and extraction calls; a move from Scrapy Cloud means arranging another runtime for the crawler. Keep extraction logic, schemas and raw outputs in systems your team controls where possible, and confirm warehouse formats and retention terms for a managed-data project.

The API is designed to handle access strategies, browser rendering and structured extraction behind one request surface. That can reduce crawler infrastructure work, but it does not remove the need to validate the data: check that the requested fields are present and meaningful, rather than treating an HTTP success as a successful scrape. The vendor says the API bills on success; confirm how success is defined for your target sites and workload.

Zyte does not publish rate limits on its homepage, so ask about concurrency, throughput, retry behavior and account-level limits before setting production targets. Also review how API keys are managed, where traffic and response data are processed, and the compliance controls relevant to your sources. Reddit commenters warn that site mix and JavaScript requirements can change costs, and that status 200 alone may not mean a usable product parse. One commenter describes Zyte as an all-in-one scraping option, but the discussions do not establish comparative reliability. Test against your own domains and data-quality checks.