Check it before you trust it.
Paste a package, a script or a website. We run it in a throwaway sandbox where it can't hurt anything, then an analyst tells you what it did and whether it's safe. Verdict first, evidence underneath.
- Isolation
- Sandbox per job
- Retention
- 7 days, then gone
- Price
- From $1 / check
“Is this site in good shape?”
Website health
A real browser visits the site from the sandbox and checks what people, regulators and AI assistants will see.
6 agents
Launch Check
Everything to check before you hand a site to a client, in one report.
- Input
- Website URL
Website Auditor
A clear, prioritized health check of any website, as a PDF.
- Input
- Website URL
Migration Check
Before you switch to a new site: will every old link, ranking and bookmark still land somewhere?
- Input
- Current site + new site
Accessibility Checker
WCAG 2.2 check of your site in a real browser, with fixes anyone can follow.
- Input
- Website URL
Cookie Consent Checker
Does your site track visitors before they agree, and does 'reject' really reject?
- Input
- Website URL
Form Tester
Does your contact form actually deliver? We send a real test message and check.
- Input
- Page with a form
“How do we compare?”
Marketing & competition
See what AI assistants recommend and how your site stacks up against competitors, with concrete changes to make.
2 agents
“Is this safe to trust?”
Safety & trust
Packages, scripts and suppliers checked before they get near your data. Code runs in a sealed sandbox; an analyst explains what it found.
3 agents
Vendor Security Snapshot
Before you trust a supplier with your data: what their public footprint says about security and privacy.
- Input
- Website URL
Package Check
Check an npm or PyPI package for malware before you install it.
- Input
- npm / PyPI package
Script Inspector
Find out what a script really does before you run it.
- Input
- Script file or text
Experiments
Labs
Useful tools that don’t fit the “check it” story yet. Expect rough edges.
1 agent
How it works
A lab and an analyst, for about a minute
- 01
Paste it, or pick an agent
A package name, a script, a URL. Plus your email so the report can find you.
- 02
It runs in a sealed sandbox
No access to your machine, your network or our secrets. Every file, process and connection is recorded.
- 03
An analyst writes the verdict
The AI reads the evidence and explains it in plain words: what happened, why it matters, what to do. Then the sandbox is destroyed.
Every report, same shape
Verdict first. Then why. Then the raw evidence.
Free scanners give you a log. We give you a decision you can act on, and keep the log underneath for when you want to check our work.
Why “micro environment”
The thing you're checking never touches you.
Every job gets its own sandbox that exists only while the agent works. Foreign code and files never touch anything else, and nothing lingers afterwards.
Read what we keep- Isolation
- One sandbox per task. Foreign code and files touch nothing else.
- Safety
- Unknown scripts run where they can do no harm: no network, no secrets.
- Privacy
- The environment is wiped after the job; results are kept 7 days at most.
- Predictable price
- Every task has a time and step limit, so the cost per task is capped.
Something else you'd like checked before trusting it?
Browser extensions, Docker images, lockfiles, suspicious emails. Tell us what you'd use.