Security & data handling
Last updated:
This page answers the questions in Figma's plugin data-security review and what security-minded teams usually ask us.
What data is sent to the service?
- The page address, chosen screens and options, and an account identifier derived from your Figma user id (no name, no email).
- A license key or a short-lived Figma purchase token, if you have a paid plan.
- In Local file mode: the files you upload. In AI mode: the HTML your AI tool sends. With Send to Figma: the capture of the tab you chose.
Does the plugin read my Figma file?
No. The plugin writes layers into your file. It reads only the information it stored on its own frames (source address and width) to offer Re-import. Your fonts are checked locally and never uploaded. Your file's content is never sent anywhere.
How long is data kept?
| Data | Kept |
|---|---|
| Import result | Delivered once, deleted from memory; job record 15 minutes |
| Uploaded files | Memory only; deleted right after the last page (2 minutes if abandoned) |
| Send to Figma / AI inbox | Memory only; deleted when your plugin takes it, at most 30 minutes |
| Import log | 180 days; IP address removed after 30 days |
| Security log | 90 days |
| Free-plan IP counters | 2 months |
Page content is never written to disk.
Encryption
All traffic between the plugin, the extension and the service uses HTTPS (TLS). AI and extension connection tokens are stored only as SHA-256 hashes.
Isolation
The service runs as three separate containers with their own permissions:
- API — plans, limits and licenses. It never runs a browser or decodes images.
- Browser — a sandboxed Chromium with a strict seccomp profile, no volumes and no secrets, reachable only from an internal network.
- Image decoder — converts downloaded images with no internet access and no way to reach the other containers.
All three run with read-only file systems, all Linux capabilities dropped and no privilege escalation. The browser is restarted regularly with a clean temporary disk.
Network protection
- A firewall lets the browser reach only public web ports; private networks, cloud metadata addresses and internal services are blocked.
- SSRF protection checks the address you enter, every request the page makes and every image download, including redirects; DNS is pinned at connection time.
- Uploaded files are opened from memory on a virtual address — never through
file://.
Abuse prevention
Rate limits apply per IP address and per account. Clear attack patterns — scanning for secrets, internal addresses, malformed requests — lead to an automatic ban of both IP and account identifier, escalating from 24 hours to 30 days.
Third parties
None for your content. We use a hosting provider for servers and Cloudflare to deliver this website. Figma verifies purchase tokens for Pro.
Reporting a vulnerability
Email [email protected] with the subject “Security”. Please give us reasonable time to fix an issue before disclosing it.