# Framejet Screenshot

> Clean PNG/JPEG screenshots via REST or MCP, with goal-driven multi-step navigation.

- name: dev.framejet/screenshot
- version: 1.1.0
- connection class: R1 (autonomous)
- trust: 83/100 flags: no-repository
- quality: 83/100 (good)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/dev.framejet%2Fscreenshot
- tools: https://api.protogrid.dev/v1/servers/dev.framejet%2Fscreenshot/tools

**An agent can connect on its own if it already holds `SCREENSHOT_TOKEN`.** Remote endpoint behind a static API key or header. The registry never sees the value; the agent fills the placeholder.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "screenshot": {
      "type": "http",
      "url": "https://framejet.dev/mcp",
      "headers": {
        "Authorization": "Bearer ${SCREENSHOT_TOKEN}"
      }
    }
  }
}
```

Placeholders to fill: `${SCREENSHOT_TOKEN}`.

## Trust

- hygiene: 80
- liveness: 100
- freshness: 100
- provenance: 55
- drivers: +dns-namespace +website +reachable +uptime-100% +updated-29d-ago -no-repository
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 83/100 (good)

### protocol: 88 (weight 25)

- pass `protocol.modern`: supports 2026-07-28 (server/discover)
- pass `protocol.stateless`: answers without a session
- pass `protocol.transport`: streamable HTTP
- warn `protocol.list_ttl`: tools/list declares ttlMs 0, so clients re-fetch it on every use

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 71 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- fail `tools.annotations`: no tool declares readOnlyHint or destructiveHint, so clients cannot tell safe calls from risky ones
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: screenshot (readOnlyHint, destructiveHint, idempotentHint, openWorldHint)
- pass `tools.token_cost`: about 703 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 100 (weight 15)

- pass `stability.changes`: 1 tool changes in 30 days
- pass `stability.rug_pull`: no tool changed its meaning under the same name

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 1, about 703 tokens to load them all
- tool set hash: f266e2e545048dd25bef97092e893eb1
- tools last changed: 2026-09-30T10:41:53.060Z
- badge: [![protogrid quality](https://protogrid.dev/badge/dev.framejet/screenshot.svg)](https://protogrid.dev/servers/dev.framejet/screenshot)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

- 2026-09-30 `screenshot` changed (description, inputSchema)
  - before: Render any public URL to a PNG or JPEG image. Cookie/consent banners and chat widgets are removed by default so the result is clean enough to reason about. Returns the image inline. For a page that is several clicks past the URL, pass `goal` in plain words and Framejet walks the flow before capturing — you do not need CSS selectors, and you could not write them anyway, because the controls for step 2 do not exist until step 1 has happened. Any text to be typed must be supplied in `values`; nothing is ever invented. If the goal cannot be reached the call fails and no quota is spent, so a returned image always means the goal was reached. When you do know the page, `actions` is cheaper, faster and deterministic.
  - after: Render any public URL to a PNG or JPEG image. Cookie/consent banners and chat widgets are removed by default so the result is clean enough to reason about. Returns the image inline. For a page that is several clicks past the URL, pass `goal` in plain words and Framejet walks the flow before capturing — you do not need CSS selectors, and you could not write them anyway, because the controls for step 2 do not exist until step 1 has happened. Any text to be typed must be supplied in `values`; nothing is ever invented. If the goal cannot be reached the call fails and no credits are spent, so a returned image always means the goal was reached. On plans that price goals per block (10 credits a block, a screenshot is 1) `max_credits` is the most a goal may spend; it stops with goal_budget_exceeded rather than go past it. When you do know the page, `actions` is cheaper, faster and deterministic.

Full history: https://api.protogrid.dev/v1/servers/dev.framejet%2Fscreenshot/changes

## Remotes

- https://framejet.dev/mcp (streamable-http, auth api_key ${SCREENSHOT_TOKEN}, reachable true, uptime30d 1)

## Packages

_none_

## Tools (1)

- `screenshot`: Render any public URL to a PNG or JPEG image. Cookie/consent banners and chat widgets are removed by default so the result is clean enough to reason about. Returns the image inline. For a page that is several clicks past the URL, pass `goal` in plain words and Framejet walks the flow before capturing — you do not need CSS selectors, and you could not write them anyway, because the controls for step 2 do not exist until step 1 has happened. Any text to be typed must be supplied in `values`; nothing is ever invented. If the goal cannot be reached the call fails and no credits are spent, so a returned image always means the goal was reached. On plans that price goals per block (10 credits a block, a screenshot is 1) `max_credits` is the most a goal may spend; it stops with goal_budget_exceeded rather than go past it. When you do know the page, `actions` is cheaper, faster and deterministic.