How does Briefwell keep one customer's data separate from another's?

Every database table enforces row-level security so a workspace can only read its own data, and this tenant isolation is verified continuously in CI, which proves zero cross-tenant reads.

How are credentials and API keys stored?

Connector credentials, channel secrets, and integration tokens are encrypted with AES-256-GCM with the key kept outside the database. API keys are hashed at rest and shown only once, at creation.

Does Briefwell use ChatGPT or Claude subscriptions to run its AI?

No. Model calls use only the official OpenAI and Anthropic APIs with server-side API keys — never a subscription session, browser credential, local proxy, or compatibility bridge.

Does Briefwell access content behind logins or paywalls?

No. Briefwell monitors public pages only and never bypasses a login, paywall, or CAPTCHA. It does not use anyone's subscription session or browser credentials to reach gated content.

Security and Privacy at Briefwell

Briefwell's security posture covers tenant isolation, encryption of stored credentials and secrets, use of official AI provider APIs only, and monitoring that stays on public pages.

Briefwell's security posture rests on tenant isolation, encryption of stored credentials and secrets, use of official AI provider APIs only, and monitoring of public pages exclusively. Every database table enforces row-level security, verified continuously in CI to prove zero cross-tenant reads. Model calls use only the official OpenAI and Anthropic APIs with server-side keys, and Briefwell never bypasses a login, paywall, or CAPTCHA to reach a source.

Start free Book a walkthrough Watch the demo ↗

At a glance

Tenant isolation: Every table enforces row-level security; CI proves zero cross-tenant reads on each run.
Encryption: Connector credentials and channel secrets are encrypted with AES-256-GCM, keys kept outside the database.
API keys: Hashed at rest and shown only once at creation; the plaintext is not retained.
Official APIs only: Model calls use the official OpenAI and Anthropic APIs with server-side keys — no subscription sessions or proxies.
Public pages only: Briefwell monitors public pages and never bypasses logins, paywalls, or CAPTCHAs.
Read-only connectors: Client data connectors to warehouses and CRMs are read-only.

Tenant isolation

Every table in Briefwell's database enforces row-level security, so one workspace's data is not readable by another. This is verified in continuous integration, which proves zero cross-tenant reads as part of the test suite rather than relying on isolation being correct by convention alone.

Encryption of credentials and secrets

Connector credentials, channel secrets, and integration tokens are encrypted with AES-256-GCM, with the encryption key never stored in the database itself. API keys are hashed at rest and are reveal-once at creation — Briefwell does not retain a copy of the plaintext key after it is shown.

AI model calls and web access

Model calls use only the official OpenAI and Anthropic APIs with server-side API keys — never a ChatGPT or Claude subscription session, browser credential, local proxy, or third-party compatibility bridge. Briefwell monitors public pages only and never bypasses a login or paywall to reach a source, and it does not solve CAPTCHAs or otherwise defeat bot detection. Web fetching includes defense in depth against server-side request forgery.

Your own data stays read-only

When a workspace connects its own data — a CSV or spreadsheet uploaded as a queryable dataset, or a read-only connector to a warehouse or CRM — Briefwell reads from those sources but does not write back to them. Connected credentials are encrypted and scoped to that workspace, and the tenant isolation that separates one customer's data from another's applies to connected data the same way it applies to monitoring results.

How it works

Point Briefwell at the web pages, feeds, APIs, or research topics you want to watch.
It checks them on a schedule, compares each result with the previous version, and keeps a record of what changed.
When a change matters, Briefwell creates an update — judged and explained, with the evidence attached — and delivers it to Slack, email, your CRM, or a signed webhook.

Frequently asked questions

How does Briefwell keep one customer's data separate from another's?: Every database table enforces row-level security so a workspace can only read its own data, and this tenant isolation is verified continuously in CI, which proves zero cross-tenant reads.
How are credentials and API keys stored?: Connector credentials, channel secrets, and integration tokens are encrypted with AES-256-GCM with the key kept outside the database. API keys are hashed at rest and shown only once, at creation.
Does Briefwell use ChatGPT or Claude subscriptions to run its AI?: No. Model calls use only the official OpenAI and Anthropic APIs with server-side API keys — never a subscription session, browser credential, local proxy, or compatibility bridge.
Does Briefwell access content behind logins or paywalls?: No. Briefwell monitors public pages only and never bypasses a login, paywall, or CAPTCHA. It does not use anyone's subscription session or browser credentials to reach gated content.