Security and reliability

ChannelForge publishes as you on the platforms you connect. This is how it protects your accounts, your channels and your content — and how it keeps publishing when something breaks.

Network and transport

Traffic is encrypted all the way, and the servers can only be reached through Cloudflare.

TLS end to end: from your browser to Cloudflare, and from Cloudflare to our origin with a Cloudflare origin certificate.
The origin accepts HTTPS only from Cloudflare’s IP ranges, so it can’t be reached around the edge.
HTTP Strict Transport Security (HSTS) keeps browsers on HTTPS.
The app sends a strict Content-Security-Policy with a fresh nonce on every request, and frame-ancestors none, so no other site can frame it.

Accounts and sessions

Sign-in is protected in layers: strong hashing, a second factor and sessions you control.

Passwords are hashed with argon2id using OWASP-recommended parameters.
Two-factor authentication with any TOTP authenticator app, plus one-time recovery codes.
Sessions are stored server-side as hashes, rotated on every sign-in and revocable from your settings.
Sign-in, signup and the API are rate limited.

Access control

People see what their role allows, and workspaces never see each other.

Role-based access in every workspace: owner, admin, editor and viewer.
Every database query is scoped to the workspace it belongs to.
CSRF tokens and Origin checks on every change.
An audit log records security- and billing-relevant actions.

Data protection

Secrets are sealed, storage is private, and the service is hosted in the EU.

OAuth tokens for the platforms you connect, and every other secret, are sealed with AES-256-GCM at rest.
The database (AWS Aurora) is encrypted at rest and runs on a private network, with automated backups and point-in-time recovery over the last 7 days.
Media lives in a private S3 bucket and is served only through short-lived signed URLs.
Hosted in AWS eu-central-1 (Frankfurt), behind Cloudflare.

Isolated generation

The machines that render your episodes are given as little as possible.

Each generation job runs in its own isolated workspace, under its own unprivileged user, on the GPU machine.
Jobs never hold our cloud or model credentials.
Model calls go through our metering gateway with a per-job, budget-capped token, so a job can’t spend beyond its budget.

Payments

Card details never touch our servers.

Payments are handled by Stripe and PayPal on their own pages. We never see or store card numbers.

Reliability

Publishing keeps going when a platform or a job has a bad moment.

A self-healing scheduler watches every publishing slot and catches up any slot that was missed.
Failed uploads are retried safely: a post is never published twice.
GPU work is queued with fair-share scheduling, and every job reports live progress.
Episodes that fail a quality gate go back for fix rounds instead of going out.

Reporting a vulnerability

If you believe you’ve found a security issue, email support@storeswiki.com with what you found and the steps to reproduce it. Please test only against your own workspace, don’t access or change data that isn’t yours, and give us a reasonable chance to fix the issue before you share it publicly. We’ll confirm we received your report and keep you updated while we work on it.