Velira can never move your money.

Velira holds some of your most sensitive information — the full picture of what you own, owe, earn, and plan. We engineer the product so that protecting it isn't a policy bolted on afterward; it's built into how the system works. This page describes the safeguards that are actually in the running product today. As an early-access company we are still maturing our formal program, so we describe how we work and what we're building toward — never a certification or audit we have not yet earned.

The single most important fact: Velira can never move your money

When you link a bank or brokerage, the connection runs through a regulated third-party data aggregator (Plaid for banks, SnapTrade for brokerages) and is strictly read-only. Velira has no ability to move money, place a trade, or change anything in your accounts — there is no such function in the product at all. This isn't a setting we promise not to use; the capability does not exist in our code. Even in an absolute worst case, the most anyone could see is information — never the keys to your money. Every recommendation Velira makes is something you choose to act on, in your own brokerage.

Your data is isolated from every other customer — enforced by the database

Multi-tenant systems usually rely on application code to keep one customer's data from reaching another. Velira does that andenforces it one layer deeper, at the database itself: every customer's records are protected by row-level security (RLS), so a query physically cannot return another customer's rows — even if a bug in feature code forgot to filter. There is no "read all customers" path in the system by design. This isolation is covered by a permanent automated testthat fails the build if cross-customer access ever became possible, so the guarantee can't silently regress.

Encryption in transit and at rest

All data moving between your device and Velira is protected with industry-standard TLS, enforced site-wide. Data we store lives on managed, access-controlled infrastructure that encrypts it at rest. For especially sensitive records we additionally apply per-customer application-layer envelope encryption, where each customer's sensitive data is encrypted under its own key with AES-256-GCM — which also enables cryptographic erasure (see deletion below). Together these reduce the risk of your information being intercepted or exposed.

Protected, least-privilege access

Access to your account is protected by authentication, with multi-factor authentication available through our identity provider. Repeated failed sign-in attempts are rate-limited to slow credential-stuffing attacks. Internally we follow the principle of least privilege: access to systems and customer data is limited to what is strictly needed to operate and support the service, granted to as few people as practical, and kept small as we grow. Secrets and API keys live in managed environment configuration — never in our source code.

Monitoring, logging, and abuse protection

We use continuous error monitoring with personal information scrubbed from reports, so diagnostics never leak your data. Any privileged action that views or changes customer data is recorded in an audit log. Inbound webhooks from our providers are cryptographically signature-verifiedbefore we act on them, and money-adjacent and AI endpoints carry rate limits and spend ceilings so they can't be abused or run away.

You own and control your data

You can export everything we hold for you at any time as a portable copy, and you can request deletion of your account and personal data, which we erase as described in our Privacy Policy — including cryptographic erasure where envelope encryption is in use, so the data becomes unrecoverable. We practice data minimization — we collect only what the product needs to do its job. And to be unambiguous: we never sell your data, and we never share it for advertising.

Who we rely on to operate

Running Velira responsibly means using a small set of vetted providers, each for a specific purpose and each receiving only the data required for that purpose:

• Plaid— read-only bank & transaction connections.
• SnapTrade — read-only brokerage holdings.
• Vercel (hosting, database, storage) — application infrastructure.
• Upstash — operational cache and the job queue behind background syncs and briefings.
• Anthropic — the AI coach; grounded on your data only to answer your questions, not used to train models.
• Clerk — sign-in and account identity.
• Resend — transactional and briefing email.
• Sentry — error monitoring, with personal information scrubbed from reports.
• Stripe — payments and merchant of record (when billing is enabled), so we never handle raw card numbers ourselves.
• Cartesia — speech-to-text and text-to-speech for voice features (only when voice features are enabled).

We work to keep this list current and to hold each provider to data- protection terms appropriate to the data they touch.

What we're still building

We believe in being straight about where we are. Today the safeguards above are live in the product. We have not yet completed a formal third-party certification (such as SOC 2) or an independent penetration test — those are on our roadmap, and we will say so plainly here when they're done rather than imply them before. If a formal control matters to your decision, ask us where it stands and we'll tell you honestly.

Responsible disclosure

Security is an ongoing effort, and we welcome help from the community. If you believe you've found a vulnerability, please report it to security@velirafinance.com before disclosing it publicly, give us a reasonable window to investigate and fix it, and avoid accessing or modifying data that isn't your own. We appreciate good-faith research and will work with you in good faith in return.

Contact

Questions about security, our practices, or a report to make? Email security@velirafinance.com.