Desktop and mobile website screenshots plus page context for AI agents and automation workflows.
Scored across 2 tools
capture_website produces visual screenshots, while read_page_context returns semantic page content like text, links, and forms. Their descriptions clearly separate visual capture from content extraction, leaving little room for misselection.
Both tools follow a clear verb_noun pattern: capture_website and read_page_context. The naming is consistent, readable, and reflects each tool's action and target.
With only two tools, the server feels thin but not unreasonable for a focused web-page capture and reading utility. Each tool serves a distinct purpose, though the surface is minimal.
The server covers the core needs of capturing a website visually and extracting readable page context. Minor gaps exist, such as no element-specific capture or screenshot management, but the main workflows are functional.
Capture desktop, mobile, or both screenshots for a website. Successful screenshots are reused from cache until TTL expiry; failed or blocked captures are never cached and are retried live on the next call.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | Website URL or domain to capture. | |
| device | No | Screenshot device variant. | both |
| delayMs | No | Extra settling delay before capture. | |
| fullPage | No | Capture the full scrollable page instead of only the viewport. | |
| ttlSeconds | No | Cache TTL in seconds. 0 disables reuse for this request. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already indicate read-only, idempotent, and non-destructive behavior. The description adds valuable behavioral context: successful screenshots are cached until TTL expiry, while failed or blocked captures are never cached and are retried live on the next call. This goes beyond the annotations and does not contradict them.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two concise, front-loaded sentences. The first states the action and scope, and the second adds the essential cache and retry behavior. Every sentence earns its place with no filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a read-only capture tool with fully documented parameters and safety annotations, the description covers purpose, device variants, and cache/TTL behavior. The main gap is that with no output schema, it does not explicitly state the return format, such as image URLs or binary image data, so the agent must infer the output shape from the tool name.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema documents all five parameters with 100% coverage, so the baseline is 3. The description adds meaning by explaining the cache/TTL behavior tied to ttlSeconds and by reinforcing the desktop/mobile/both options, particularly the non-obvious fact that failed captures are never cached.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description states a specific verb and resource: capture desktop, mobile, or both screenshots for a website. It clearly indicates what the tool does, though it does not explicitly differentiate itself from the sibling tool read_page_context.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Usage is implied by the action verb and device variants: use this tool when a visual screenshot of a website is needed. However, there is no explicit guidance about when to prefer this tool over read_page_context or when not to use it.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Render a public web page and return a compact Community agent-readable representation: visible text, clean Markdown, headings, links, forms, basic image metadata and interactive elements with bounding boxes. No OCR, visual inference, crawl, sessions or advanced action selectors.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | Public website URL or domain to understand. | |
| delayMs | No | ||
| viewport | No | desktop | |
| maxTextChars | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already establish readOnly, idempotent, non-destructive, and open-world behavior. The description adds useful context beyond annotations: it emphasizes the page must be public (auth expectations), the result is a compact text representation, and sessions/selectors are unsupported. It does not mention rate limits or failure modes, but for a read-only fetch that is a minor gap.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single dense sentence that front-loads the core action and result, then enumerates outputs and exclusions efficiently. There is no filler or redundant restating of the tool name.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With no output schema and only partial parameter documentation, the description gives a good high-level view of return content but leaves parameter semantics underspecified. An agent could invoke with defaults, but understanding how delayMs, viewport, or maxTextChars affect behavior would require guessing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is only 25%, with url being the only described parameter. delayMs, viewport, and maxTextChars have types/defaults but no semantic explanation, and the description does not compensate by explaining these parameters or how they affect the rendered output.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('Render'), the resource ('a public web page'), and the concrete output ('visible text, clean Markdown, headings, links, forms, basic image metadata and interactive elements with bounding boxes'). It also differentiates itself from visual/crawling tools by explicitly excluding OCR, visual inference, crawl, sessions, and advanced action selectors.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description conveys it is for text-only agent consumption and lists what it does not do, giving clear context on appropriate use. However, it never names the sibling tool capture_website or explicitly says 'use this instead when you need X', so it stops short of full when/when-not alternative guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
capture_websiteread_page_contextCapture screenshots, detect visual regressions between page versions, and analyze with AI.
Screenshots, PDFs and Markdown from any URL or HTML for AI agents, via the SnapForge API
Screenshot, diff, audit and sitemap-capture any web page — 5 MCP tools for AI agents.
Hosted browser for AI agents: screenshots, post-JS DOM, console, WCAG. No install, no API key.
Add one secure layer between your agents and this server.
