The website QA tool for your coding agent
squirrelscan audits your website for SEO, performance, security, accessibility and agent experience issues, and gives your coding agent exact fixes.
Built for coding agents
Skills for Claude Code, Cursor, Codex and any coding agent run the audit, fix and re-audit loop. Connect over the CLI, MCP or the API.
Smart crawling
Fast and memory efficient, incremental with ETag, Last-Modified and content hashing, and resumable from checkpoints.
Enriched fixes for you and your coding agent
Every finding says how to fix it, in LLM-optimized output your agent can apply, across 300+ rules in 22 categories.
Any agent. One prompt. Zero issues.
squirrelscan plugs into the agent you already run. Tell it to fix your site: it audits, applies exact fixes, and re-runs until nothing's left. Nightly audits and a CI gate keep it that way.
Works with your agent
You, to your coding agent
A report worth sharing
Readable enough for your team, structured enough for your coding agent. The same report works for both.
Health score & grade
An at-a-glance score and letter grade, plus scores for SEO, performance, security and agents.
Fix-it guidance
Each issue ships with the concrete fix, the same detail your coding agent needs.
Self-contained HTML
One portable file. No server, no dependencies. Open it or share it anywhere.
Paste it into your agent
The fixes travel with the report. Drop it into Claude Code or Cursor and say: work through these.
Seven output formats
console · html · markdown for humans; llm · json · xml · text for agents.
A live report: squirrelscan auditing its own site, 220 pages. Click through the issues.
Four ways to run it
Same audit engine, same report, four front doors: your terminal, your coding agent, the cloud, or MCP.
CLI
A single binary, zero dependencies. Audit from your terminal and pipe results anywhere.
In your agent
An install-on-demand skill for Claude Code, Cursor, and any coding agent.
Cloud
Browser rendering, AI analysis, and shareable reports, all billed in credits.
MCP
A local stdio server (squirrel mcp) that exposes the audit engine and cloud features as agent tools.
Wiring it into CI? squirrelscan returns CI-friendly exit codes, so you can fail the build on score or severity thresholds.
300+ audit rules and counting
22 categories, and every rule comes with a severity and a fix. The website audit tool runs all of them across your site.
Accessibility
ARIA, focus, landmarks
Blocking
Ad blockers and privacy filters
Agent Experience
llms.txt, AI crawlers, LLM parsability
Analytics
GTM, consent mode
Content
Quality, readability, freshness
Core SEO
Meta tags, canonical URLs, h1, favicon
Crawl
Robots.txt, sitemap, indexing
E-E-A-T
Experience, expertise, trust signals
i18n
Hreflang, lang attribute
Images
Alt text, dimensions, formats
Site Integrity
Injected pages, phishing kits, cloaking
Legal
Privacy, cookies, TOS
Links
Internal, external, broken links
Local SEO
NAP consistency, geo meta
Mobile
Viewport, tap targets
Performance
Core Web Vitals, LCP/CLS/INP
Reliability
Broken scripts, stylesheets and forms
Schema
JSON-LD structured data
Security
HTTPS, CSP, leaked secrets
Social
Open Graph, Twitter cards
URL Structure
Length, hyphens, keywords
Video
VideoObject schema, accessibility
Rules are continuously updated. Run squirrel self update to get the latest.
Go further in the cloud
Signed out, your CLI runs a full audit locally, for free. Log in and squirrelscan enriches it in the cloud, on credits.
Browser rendering
Audit JavaScript-heavy sites and SPAs with a real headless browser.
AI page analysis
LLM parsability scoring and E-E-A-T authority signals on the pages that matter.
Technology detection
385 fingerprints: CMS, frameworks, CDNs, analytics, payments, and security.
Site profilePro
Type, audience, identity, socials, and domain age, used to keep rules relevant.
Dead-link checking
External links verified against a shared global cache.
Editor's summaryPro
A Claude-written exec summary with the big-ticket items and deltas since last run.
Publish & dashboard
Auto-publish shareable reports at reports.squirrelscan.com with score history over time.
Scheduled audits
Re-run on autopilot from the dashboard. Free schedules one site weekly, Pro goes daily on every site.
Signed out, or with --offline, local audits run free. How the cloud works
Local CLI audits are free signed out or with --offline. Signed in, audits run on credits.
A signed-in squirrel audit spends 50 credits plus 2 per audited page and adds AI analysis, dead-link checks, and published reports. Local CLI audits are free signed out or with --offline. Free includes about 5 audits a month, Pro about 30 for $19/month.
Most popular Team $29/ seat / month min 2 seats about 30 cloud audits a month per seat Upgrade to TeamFor agencies: white-label reports | |||
|---|---|---|---|
| Audits & credits | |||
| Monthly cloud credits | 500 | 3,000 | 3,000 / seat |
| Crawling & rendering | |||
| Max pages per cloud audit | 500 | 2,000 | 10,000 |
| Pages rendered at once, per audit | 1 | 5 | 10 |
| Scheduled audits | Weekly, 1 site | Daily, all sites | Daily, all sites |
| Reports & history | |||
| Score history per site | not included | included | included |
| Team & seats | |||
| Seats | 1 | 1 | 2 minimum |
| Pooled org credits | not included | not included | included |
| White-label reports | not included | not included | included |
- Failed audits refunded automatically
- Per-audit spend caps
- No card needed for Free
- Cancel anytime
Enterprise: contact us
Custom contracts and invoiced billing for larger teams.
A typical 25-page audit costs 100 credits. See what credits buy
From the nutshell
The knowledge base: generative engine optimization, AI crawlers, MCP servers, and how squirrelscan compares to the crawlers you already use.
Best generative engine optimization tools: trackers, site audits and content tools
Fifteen GEO tools sorted by the job they do, tracking whether AI engines mention you, auditing your site so they can reach it, or rewriting content so they might, with prices we read on each vendor's own page.
Best SEO MCP servers: 11 checked, with prices and setup
Eleven SEO MCP servers sorted by the job they do: keyword and backlink data, your own Google data, or a crawl of your site. Price, sign-in and scope read from each vendor's own docs.
Screaming Frog alternatives: free, open source and paid, with prices
The free and open source crawlers first, then the paid desktop and cloud ones, each priced from the vendor's own page, and the case for a CLI that runs in CI instead of a GUI you open once a quarter.
Generative engine optimization checker: test your whole site, not one URL
Run a generative engine optimization check across your whole site from the terminal, see how six free checkers did on the same site, and get the rule-by-rule checklist with a CI gate.
From the blog
Updates, guides, and thoughts on website auditing, SEO, and building tools for AI workflows.
Turns out detecting broken links is haaaard. Meet squirrel perch.
Our old link checker cried wolf on links that worked fine. We rebuilt link checking for a web that hates robots, so a broken link in your report is far more likely to be really broken.
Introducing the entity map: your site's structured data as one graph
Every squirrelscan audit now collapses your site's JSON-LD into one graph, so you can see whether you declare one organization or a new one on every page, and hand the fix to your coding agent.
2.8x performance improvement and larger audits with squirrel
A 2,500-page audit went from 502 seconds to 177. Local audits go to 10,000 pages. Here are the measurements, the conditions they were taken under, and what actually got faster.