siteglassSAFE
Archive any URL to a shareable snapshot (rrweb/pdf/png) + agent-driven web QA — an MCP server for siteglass.io. No CAPTCHA, no signup.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
`siteglass-mcp` exposes siteglass.io as MCP tools. Three things an AI agent (or a human) can do with it:
- Scan the app you're building — FREE, no account —
siteglass_scan_local
runs a headless browser on your machine against http://localhost:PORT (or any URL you can reach — nothing is tunneled to us), then siteglass analyzes the crawl and returns real findings: broken links/assets, console errors, accessibility & SEO issues, forms worth testing, plus a plain-English report. Point it at your dev server right after the agent builds or changes the app. No signup, no ownership verification, no payment. (First run downloads a headless Chromium, ~1 min.)
- Archive any public URL —
siteglass_renderrenders a page server-side
and returns a permanent, shareable snapshot: an interactive rrweb DOM capture (post-JS, selectable text, working links), a single-page PDF, or a full-page PNG. No CAPTCHA, no signup, no ownership check, callable programmatically — a bot-friendly alternative to archive.today. (It renders JS-heavy pages; it does not produce WARC, and if the target itself blocks datacenter IPs the capture can come back empty.)
- Web-QA a deployed site you own — **register → verify → scan → generate
flows → run → read results**, autonomously, no human signup. This is the hosted/continuous tier (runs on our servers against your live site).
The local scan (1) is free and needs no auth. For the hosted tools (2, 3), auth is automatic (creates + caches an API key at ~/.siteglass/key; set SITEGLASS_API_KEY to use your own); the first scan per site is free, then scans (2 credits) and flow runs (1 credit) draw from a prepaid balance topped up in USD over Bitcoin Lightning.
Install
npx siteglass-mcp
Claude Code:
claude mcp add siteglass -- npx siteglass-mcp
**Cursor
5a21de1e29b6OBSERVED · 2026-10-11Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add siteglass-mcp --env SITEGLASS_API_KEY=${SITEGLASS_API_KEY} -- npx -y [email protected]Exposed tools (12)
8 read · 4 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
siteglass_feedback | write | Send feedback to the siteglass team (bugs, missing features, anything). Works for agents and people alike. |
siteglass_generate_flows | read | Derive executable E2E test flows from a scan (LLM). Step 4. Returns flow ids. |
siteglass_get_render | read | Fetch an archive once its Lightning invoice is paid. Returns the permalink when ready, or a pending status. |
siteglass_get_run | write | Read a flow run |
siteglass_get_scan | read | Read a scan |
siteglass_register_site | read | Register a web app (its deployed URL) with siteglass and get an ownership-proof token to place on it. Step 1. |
siteglass_render | read | Archive ANY public web page for a few sats — a permanent, shareable snapshot, returned as a permalink. The bot-friendly alternative to archive.today: no CAPTCHA, no signup, no ownership verification, callable programmatically. Default |
siteglass_run_flow | write | Run one test flow. Returns a run_id; if still running, poll with siteglass_get_run. Step 5. |
siteglass_scan | read | Crawl a verified site (first scan per site is free). Returns a crawl_id; if still running, poll with siteglass_get_scan. Step 3. |
siteglass_scan_local | read | Scan a web app you |
siteglass_set_credentials | write | Provide a test account (username/password) so authenticated flows can log in to the site. |
siteglass_verify_site | read | Check that the ownership-proof token is live on the site. Step 2 (after placing the token). |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- declared (3 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (1)
@modelcontextprotocol/sdk, playwright
Gates applied: no_behavioural_pass.
5a21de1e29b6full audit observations/trust-audit/mcp-server/siteglassio__siteglass-mcp.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-11 | 5a21de1e29b6 | SAFE | B | 89 | first audit |
Questions
What is the siteglass MCP server?
Archive any URL to a shareable snapshot (rrweb/pdf/png) + agent-driven web QA — an MCP server for siteglass.io. No CAPTCHA, no signup.
What tools does siteglass expose?
12 in total: 8 read-only, 4 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is siteglass safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does siteglass need?
It reads SITEGLASS_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does siteglass run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as siteglass-mcp at 0.2.0.
How current is this page?
The grade is for one exact copy of the source (5a21de1e29b6), read on 2026-10-11. The repository is watched and re-audited when it changes.