shield Security & Trust

Built to be trusted, not just believed

Your data is isolated and encrypted in depth, the platform's institutional exposure data carries its receipts, and every AI surface runs under a strict safety guardrail. Here is exactly how.

Three layers, not one

Sensitive data is protected at rest, isolated per user by the database engine, and encrypted at the column level, so a gap in any single layer is not a breach.

Layer 1 · At rest
storage

Transparent Data Encryption

The entire database is encrypted at rest with Azure SQL TDE, so the files on disk and every backup are unreadable without the managed keys.

Layer 2 · Isolation
group_off

Per-user query filters

Every user-owned record is scoped to its owner by database-enforced global query filters, so one account's data is never returned to another, by construction.

Layer 3 · Field level
lock

AES-256 column encryption

The most sensitive fields are additionally encrypted at the column level with AES-256, so they stay protected even inside an authorized query.

verified_user Defense in depth

Every number carries its receipt

The platform's institutional exposure data is extracted from companies' own SEC filings. We don't ask you to trust a model's guess: each figure is tied to the sentence it came from, reconciled against the filing's own XBRL totals, and cleared by a human review gate before it ever reaches you.

  • format_quote The verbatim filing quote each figure came from
  • calculate An arithmetic check against the company's XBRL total
  • how_to_reg A human review gate: nothing surfaces until it's approved
  • policy A coverage caveat on every view, so it never overclaims
disclosureRevenue by geography
reconciled✓ to the dollar
source10-K · Item 2
review✓ approved
"...revenues attributable to customers located outside the United States were..."

Guardrailed by design

Every AI surface on the platform runs under one shared safety system, and the app, not the model, controls what you're told.

verified

One shared guardrail

Every AI call runs under a single shared safety system prompt. It's informational and educational, and it never tells you to buy, sell, or allocate.

receipt_long

Accountable by default

Every AI call is logged with its feature and cost, so usage is auditable rather than a black box, and premium models are used where they're justified.

gavel

The app writes the disclaimer

The model never writes its own "this is not advice" footer. The platform attaches a standardized educational disclaimer, so the wording is controlled and consistent.

Your data, your control

download

Delete your account, export on request

You can delete your account yourself from your settings, and request a machine-readable export of your data. Your data is never sold or shared with third parties.

cloud_done

Managed-identity infrastructure

Secrets live in a managed key vault, not in code, and services authenticate with managed identities on enterprise cloud infrastructure. Market data is cross-verified across multiple providers.

info

AlgoForgeX is for educational and informational purposes only and does not constitute financial, investment, or trading advice. Everything the platform produces is designed to inform your own research, not to direct it. Always do your own research and consult a qualified, licensed financial advisor before making any investment decisions.

Serious about your stack? So are we.

Explore the platform, or reach out with a security or compliance question.