Zyte API coverage
How much of the Zyte API xyte implements, field by field, from the OpenAPI spec in spec/zyte-openapi.json. A test fails when the spec has an entry not listed here.
✅ supported · 🟡 partial · 🔌 pluggable (bring an extractor) · ⚪ accepted, no effect · ❌ not supported (request fields: rejected with a 400)
| Area | Supported or partial |
|---|---|
| Request fields | 31/56 (+25 pluggable) |
| Response fields | 13/25 (+12 pluggable) |
| Browser actions | 18/20 |
Request fields
A request is an HTTP request (httpResponseBody, httpResponseHeaders, or extraction from HTTP) or a browser request (browserHtml, screenshot, or extraction from the browser). Applies to says which kind each field belongs to: a field of the other kind is a 422 /request/unprocessable, as on Zyte.
| Name | Status | Applies to | Notes |
|---|---|---|---|
url | ✅ supported | both | IP-address hosts are accepted too (Zyte rejects them). |
requestHeaders | ✅ supported | both | referer, on HTTP and browser requests. |
tags | ✅ supported | both | No Stats API to send them to: they label the onRequest / onAttempt reports and request logs instead, for your own analytics. |
ipType | 🟡 partial | both | Reaches the provider as ctx.ipType; byIpType() routes on it. Presets ignore it. |
httpRequestMethod | ✅ supported | HTTP only | |
httpRequestBody | ✅ supported | HTTP only | |
httpRequestText | ✅ supported | HTTP only | |
customHttpRequestHeaders | ✅ supported | HTTP only | A Cookie header is dropped, as on Zyte: cookies go in requestCookies (authentication). On a browser request, Zyte answers 422 (checked live). |
httpResponseBody | ✅ supported | HTTP (makes it one) | Truncated at 10 MB. |
httpResponseHeaders | ✅ supported | both | HTTP and browser (main document) requests. |
browserHtml | ✅ supported | browser (makes it one) | Playwright Chromium, or your own browser.launch. |
screenshot | ✅ supported | browser (makes it one) | |
screenshotOptions | ✅ supported | browser only | format, fullPage. |
extractFrom | 🔌 pluggable | either (extractFrom) | httpResponseBody, browserHtml, or userHtml (from <field>Options.userHtml: with a local extractor there is no download at all). Unset, xyte extracts from HTTP; Zyte currently defaults to browserHtml for AI extraction, so set it to match. |
article | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
articleOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
articleList | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
articleListOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
articleNavigation | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
articleNavigationOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
forumThread | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
forumThreadOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
jobPosting | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
jobPostingOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
jobPostingNavigation | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
jobPostingNavigationOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
pageContent | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
pageContentOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
markdown | ✅ supported | either (extractFrom) | Built in and non-AI, like Zyte's: dependency-free converter with heuristic main-content detection. extractors.markdown overrides it. |
markdownOptions | ✅ supported | either (extractFrom) | scope (mainContent / fullPage), extractFrom. |
product | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
productOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
productList | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
productListOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
productNavigation | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
productNavigationOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
customAttributes | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
customAttributesOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
serp | 🔌 pluggable | either (extractFrom) | llmExtractors() with any LLM (Zyte output schema), or your own extractors.<field>; 400 without one. |
serpOptions | 🔌 pluggable | either (extractFrom) | Passed to your extractor as options; extractFrom picks HTTP or browser. |
geolocation | ✅ supported | both | Encoded into the provider credentials. With direct(), plain proxyUrl() or a template() without {country} it can't be applied: the request runs with an X-Xyte-Warning header / onWarning. |
javascript | ✅ supported | browser only | |
actions | ✅ supported | browser only | All of them; searchKeyword, setLocation and interaction through handlers or zyteApi() (see Actions). |
jobId | ✅ supported | both | Echoed back. |
echoData | ✅ supported | both | |
viewport | ✅ supported | browser only | |
followRedirect | ✅ supported | HTTP only | HTTP requests (browsers always follow, as on Zyte). |
sessionContext | ✅ supported | both | A pool per context (default 8 sessions; retired after 4 h or 3 bans, as on Zyte). In process, or shared across instances with sessions.store / XYTE_REDIS_URL. |
sessionContextParameters | 🟡 partial | both | actions run once per new session in a browser at the request URL, through the provider resolved for it (an API provider keeps its own sessions: no local init). Zyte's own actions run through actions handlers; without one, skipped with a warning, or a 400 with unsupportedActions: 'error' (no pass-overs during init). |
session | ✅ supported | both | Sticky exit IP through the provider, plus a cookie jar kept in memory per process. |
networkCapture | ✅ supported | browser only | url and resourceType filters; max 10 captures / 5 MiB. |
device | ✅ supported | both | Mobile user agent and client hints; mobile viewport in the browser. |
cookieManagement | ✅ supported | both | |
requestCookies | ✅ supported | both | |
responseCookies | ✅ supported | both | |
includeIframes | 🟡 partial | browser only | Each frame's document is written into its <iframe> element; Zyte doesn't document its exact format. |
Response fields
| Name | Status | Notes |
|---|---|---|
url | ✅ supported | Final URL after redirects. |
statusCode | ✅ supported | |
httpResponseBody | ✅ supported | |
httpResponseHeaders | ✅ supported | |
browserHtml | ✅ supported | |
markdown | ✅ supported | |
session | ✅ supported | Echoes the request session. |
screenshot | ✅ supported | |
article | 🔌 pluggable | |
articleList | 🔌 pluggable | |
articleNavigation | 🔌 pluggable | |
forumThread | 🔌 pluggable | |
jobPosting | 🔌 pluggable | |
jobPostingNavigation | 🔌 pluggable | |
pageContent | 🔌 pluggable | |
product | 🔌 pluggable | |
productList | 🔌 pluggable | |
productNavigation | 🔌 pluggable | |
customAttributes | 🔌 pluggable | |
echoData | ✅ supported | |
jobId | ✅ supported | |
actions | ✅ supported | action, elapsedTime, status, error; interactionLogs from your action handlers' log(). |
responseCookies | ✅ supported | |
networkCapture | 🟡 partial | url, statusCode, headers, httpResponseBody, filter; no request / interceptionStatus. |
serp | 🔌 pluggable |
Actions
| Name | Status | Notes |
|---|---|---|
click | ✅ supported | button, delay, waitForNavigationTimeout. |
doubleClick | ✅ supported | |
evaluate | ✅ supported | Runs as a script without awaiting its promises, like Zyte. |
goto | ✅ supported | |
hide | ✅ supported | |
hover | ✅ supported | |
interaction | 🔌 pluggable | Zyte runs scripts from its cloud IDE by id: with zyteApi() it does; elsewhere actions.interaction maps ids to your own scripts. Without one: skipped with a warning (or a 400 with unsupportedActions: 'error'). |
keyPress | ✅ supported | |
reload | ✅ supported | |
scrollBottom | ✅ supported | timeout, maxScrollDelay, maxPageHeight, maxScrollCount; scrollStep ignored. |
scrollTo | ✅ supported | |
searchKeyword | 🟡 partial | Built-in heuristic (searchBox: finds the search box, types, presses Enter); per-site handlers in actions.searchKeyword; Zyte's own with zyteApi(). |
select | ✅ supported | |
setLocation | 🔌 pluggable | Location forms differ per site: per-site handlers in actions.setLocation, Zyte's own with zyteApi(). Without one: skipped with a warning (or a 400 with unsupportedActions: 'error'). |
type | ✅ supported | |
waitForNavigation | ✅ supported | As the first action, configures the initial navigation. |
waitForRequest | ✅ supported | |
waitForResponse | ✅ supported | |
waitForSelector | ✅ supported | |
waitForTimeout | ✅ supported |
Errors
| Name | Status | Notes |
|---|---|---|
400 /request/invalid | ✅ supported | Malformed requests (url format, header syntax), plus fields xyte doesn't support. |
400 /request/invalid-json | ✅ supported | Server only. |
401 /auth/key-not-found | ✅ supported | Server with XYTE_API_KEYS. |
403 /auth/account-suspended | ⚪ accepted | No accounts. |
421 /website/domain-unreachable | ✅ supported | |
422 /request/unprocessable | ✅ supported | Incompatible parameters (browserHtml + httpResponseBody, browser-only fields on HTTP…) and requests asking for nothing, as on api.zyte.com. |
429 /limits/* | ⚪ accepted | No quotas; concurrency queues instead. |
451 /download/domain-forbidden | ⚪ accepted | Never emitted. |
500 /server/timed-out | 🟡 partial | Attempt timeouts are retried, then reported as 520. |
503 /limits/over-global-limit | ⚪ accepted | Never emitted. |
520 /download/temporary-error | ✅ supported | Proxy failures and network errors after retries. |
520 /download/website-ban | ✅ supported | Bans (403/429/503, challenge pages) after retries. Checked against api.zyte.com. |
521 /download/internal-error | ⚪ accepted | Never emitted. |
Beyond /v1/extract
| Name | Status | Notes |
|---|---|---|
HTTP API POST /v1/extract | ✅ supported | xyte serve, createHandler, toNodeHandler. |
| Zyte API as a provider | ✅ supported | zyteApi() (full fidelity, e.g. as a fallback) or zyteProxy() (proxy mode). |
Proxy mode (api.zyte.com:8011) | ✅ supported | xyte serve --proxy-port: Zyte-* headers, HTTPS with its own CA (xyte ca), Zyte-Request-ID / Zyte-Error-* headers (proxy mode). No HTTPS proxy interface (:8014). |
| CDP headless browser | ❌ unsupported | |
| Stats API | 🟡 partial | GET /api/stats and /api/v1/stats (USD), with the same parameters and shape, from the requests the server handled; p80s are approximate (histograms); domain_health is null; requested_features, extraction_type, apikey_labels filters are accepted and ignored. |
| Search API | ❌ unsupported |