Zyte MCP Server

fetch_http

Fetch a URL with a plain HTTP request through Zyte API: any HTTP method, optional request body, custom headers. Returns the page as Markdown by default; see 'format'. Fast and cheap; never executes JavaScript or renders the page. Use for static pages, form posts, and — with 'format': 'raw' — JSON/GraphQL/XHR API endpoints. For JavaScript-rendered pages, browser actions, or screenshots use fetch_page; for structured data (products, articles, jobs) use extract_from_http, or extract_from_browser when the page needs rendering. Returns a JSON metadata block (final URL, target HTTP status, optional headers/cookies) followed by the content in the chosen format.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`url`	Yes	Absolute http(s) URL to fetch (max 8192 chars). The host must be a domain name, not an IP address.
`tags`	No	Arbitrary key/value pairs (string values or null) attached to the request for filtering in the Zyte Stats API.
`device`	No	Device type to emulate. Default desktop.
`format`	No	How to return the response: 'markdown' (default; the main content as Markdown, without navigation, headers, footers and sidebars), 'fullPageMarkdown' (the whole page as Markdown) or 'raw' (the response body as received — text, or an image block for images). In the Markdown formats, JSON and other non-HTML text come back in a code block, and a PDF, image or other binary response as a short note; use 'raw' for API endpoints, images, or when the exact bytes matter. HEAD requests always use 'raw'.	markdown
`ipType`	No	Type of IP address to send the request from. Default: Zyte API picks the type that avoids bans for the target site.
`method`	No	HTTP method. Default GET.
`headers`	No	Custom HTTP request headers (max 200). The Cookie header is not allowed — use requestCookies. Zyte API sends some headers automatically for ban avoidance and may override or drop custom ones.
`bodyText`	No	UTF-8 text to send as the request body (max 400000 chars), e.g. a JSON or GraphQL payload. Mutually exclusive with bodyBase64.
`sessionId`	No	Client-managed session ID — a version 4 UUID you generate. Requests with the same ID reuse the same session (IP, cookies). Sessions expire 15 minutes after creation, after 2 idle minutes, or after 3 bans.
`bodyBase64`	No	Base64-encoded bytes to send as the request body (max 400000 chars encoded), for binary or non-UTF-8 content. Mutually exclusive with bodyText.
`geolocation`	No	ISO 3166-1 alpha-2 country code to route the request from (e.g. 'US', 'DE'). Default: Zyte API picks a geolocation that avoids bans and locale surprises for the target site.
`followRedirect`	No	Whether to follow HTTP redirects. Default true.
`organizationId`	Yes	Required. The Zyte organization to attribute this call to (max 100 characters, printable ASCII without spaces). If you do not already have an id, call the user_info tool: it lists the organizations your credential belongs to. IMPORTANT: if it lists more than one, ask the user which to use and wait for their answer — this call is billed to whichever organization you name here, so it is the user's choice to make, not yours. Never guess an id, and never fall back to a default. Once the user has chosen, reuse that id across the session unless they ask for a different organization.
`requestCookies`	No	Cookies to send with the request (max 100). The responseCookies output of a previous call can be passed here verbatim.
`sessionContext`	No	Server-managed session context: up to 10 name/value pairs. Zyte API reuses or creates a session per distinct context. Sessions expire after 4 hours or 3 bans.
`enableZeroTrace`	No	Keep the URL and other potentially sensitive request data out of Zyte API's request logs, metrics and stats records for this request. Default false. Use it for sensitive targets; it also leaves Zyte support less to go on when investigating the request.
`responseCookies`	No	Return the cookies set during the request in the metadata block. Prefer this over Set-Cookie response headers — it reflects the final cookie state.
`cookieManagement`	No	How cookies are handled: 'auto' (default) uses requestCookies if given, otherwise Zyte API's automatic cookies; 'discard' uses requestCookies if given, otherwise no cookies.
`verifyCertificate`	No	Validate the target site's TLS certificate and return an error instead of the page when validation fails. Default false — certificates are not validated. Browser fetches always validate them, so fetch_page has no such argument.
`sessionContextActions`	No	Browser actions run once to initialize a server-managed session for the given sessionContext (e.g. login steps). Each action takes the parameters documented in the Zyte API actions reference (https://docs.zyte.com/zyte-api/usage/reference.html), e.g. selector {type: css\|xpath, value}, timeout, source, url, keyword. Parameters are validated against the Zyte API contract.
`includeResponseHeaders`	No	Include the target response's HTTP headers in the metadata block.

Schema Changelog

Changes observed during successful MCP inspections.

No schema history has been recorded yet.

TDQS

A4.4/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden and does disclose meaningful behavior: it is fast/cheap, never executes JavaScript or renders, and returns a JSON metadata block plus content in the chosen format. It omits some operational traits (billing/rate-limit implications of organizationId are only in the schema), but the core behavioral profile is well covered.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the core purpose, then the default return, then routing guidance, then return detail. Each sentence is doing distinct work — no filler or repetition despite the dense content.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

There is no output schema, and the description compensates by describing the return shape (metadata block with final URL, status, optional headers/cookies, followed by content in the chosen format). For a 21-parameter tool it relies on the 100%-covered schema for the remaining parameters, which is reasonable, though it says nothing about the billing implications that make organizationId significant.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so every parameter is already documented in the schema. The description reiterates a few of them (method, body, headers, format) but adds no syntax or format detail beyond what the schema provides, so the baseline of 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource ('Fetch a URL with a plain HTTP request through Zyte API') and scopes it precisely (any method, optional body, custom headers). It explicitly distinguishes itself from the three most relevant siblings (fetch_page for JS rendering, extract_from_http and extract_from_browser for structured data).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-use cases (static pages, form posts, JSON/GraphQL/XHR endpoints with format:'raw') and explicit when-not with named alternatives (JS-rendered pages/screenshots -> fetch_page, structured data -> extract_from_http/extract_from_browser). Nothing is left to inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Zyte MCP Server tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?