Encryption at rest and in transit protects conversations and sensitive data.
Secure data enables smarter agents
Perfect memory requires perfect security. Every preference, constraint, and decision encrypted.
Compound learning needs clean, trusted data. Security prevents poisoned learning.
Millions in Meta, Google, TikTok campaigns. Competitor isolation is critical.
Intelligence improves over months. Long-term data security = long-term accuracy gains.
Protection built into everything we do
Strong encryption at rest and in transit. Data encrypted everywhere it moves. Agents learn from protected data.
Agents propose changes and you approve them before they run. Revoke access anytime.
Every request is checked against your role in the brand and company it reads, so one customer's data is not served to another. Pooled figures other brands see carry no brand, company or person, need at least 5 brands from 3 companies, and use only sources whose platform terms allow it.
Infrastructure monitoring across the stack.
Account-level access controls. Multi-factor authentication on roadmap.
Activity logs for account-level changes. Available in Settings.
Compliant with major privacy regulations. The application runs on Vercel and the database runs on Neon, and both hold SOC 2 Type 2 attestations. Cresva has no SOC 2 or ISO 27001 report of its own.
Export and deletion in Settings, consent before any cookie that is not essential, a public sub-processor list, and a Data Processing Agreement under Article 28.
No sale of personal data, a Your privacy choices link on every page, Global Privacy Control honoured, and deletion on request.
You control your data at all times
Clear info on what data we collect and why.
Export anytime. Delete the company from Settings.
AI agents only access data needed for insights.
Your data stays in your tenant. Not shared with other customers, not sold to third parties. See our Privacy Policy for full data-handling terms.
Choose what data to share. Opt in or out of analytics.
Delete your data whenever you want.
Quick answers about security and privacy
Your account and brand data is stored in a Neon database in Singapore (AWS ap-southeast-1), and the application runs in Singapore. Some services we use process data in the United States and elsewhere; each is named, with its region, in our subprocessor list. Data is encrypted in transit and at rest.
No. You sign in with a one-time link or a passkey, and platform connections use OAuth wherever the platform offers it. Revoke access anytime without affecting your accounts.
If we confirm a breach affecting your data, we will tell you without undue delay and no later than 72 hours after we become aware of it, with what happened, the data affected and what we are doing about it.
Yes. An owner or admin can download everything Cresva holds for the company, or for one brand, as a JSON file from Settings, Company, Export your data, at any time while the account is open.
Every request is checked against your role in the brand and company it reads, so one customer's data is not served to another. Pooled figures other brands see carry no brand, company or person, need at least 5 brands from 3 companies, and use only sources whose platform terms allow it.
Cancelling a plan does not delete your data straight away. It is deleted when an owner deletes the company in Settings, when you ask us to at hello@cresva.ai, or when the company has had no paid plan and no sign-in for 12 months, after email warnings 30 and 7 days before. A deletion an owner asks for runs 7 days after it is asked for, so it can be stopped, and is complete within 30 days. Deleted data leaves our database backups within a further 7 days.
Not yet. Sign in with a Google account, Google Workspace accounts included, or with a link sent to your email. Single sign-on through your own identity provider (SAML or OIDC, for example Okta or Microsoft Entra ID) and SCIM user provisioning are not available yet.
Internal security reviews ongoing.
What happens when a user connects a third party AI assistant to Cresva. The same wording appears in our privacy policy and terms. Last reviewed 2026-10-18.
Cresva operates a Model Context Protocol server at https://cresva.ai/mcp. A user may connect a third party AI assistant, such as Claude, ChatGPT or Cursor, to that server. The assistant is the client: it authenticates to Cresva using OAuth 2.1 and holds its own access token.
A connection can read only the brands the user selected on the Cresva consent screen at the time of granting, and only through the tools that the granted scopes permit. It cannot enumerate or reach an account the user did not grant, and it cannot reach another customer's data.
The connector is read only by default. Two tools are not reads. One adds a proposed Google campaign budget to the brand's approvals queue for a person to approve in Cresva, never changes an ad account, and needs a permission that is off unless the user grants it. The other is run_ai_shopper, which publishes a public page about a store on cresva.ai and never writes to the store.
A connection is granted by an individual user and carries that user's identity. Anything the assistant reads is what that user is entitled to read, and their permissions are re-evaluated at the time of each call rather than frozen when the grant was made. A user who loses access to a brand loses it through the connector at the same moment.
Every tool call is written to Cresva's audit log against the granting user, the assistant that made it and the grant it was made under. An administrator can therefore attribute any connector activity to a named person. Audit records are retained on the same schedule as other account audit data.
When a connected assistant calls a tool, Cresva returns the answer to that tool call to the assistant, over TLS. Nothing else is transmitted: Cresva does not send the assistant vendor credentials, connected platform tokens, other customers' data, or any bulk export of the customer's account.
When you ask a question through the Cresva Chrome extension or the npm package, the question text is sent to OpenAI to produce the answer. A Store run may send shopping questions written from the store's public product information to AI assistants such as ChatGPT, Claude, Perplexity and Gemini, to check what they recommend. When the store's products cannot be read, the questions may name the store. The Claude and ChatGPT connectors send no question text to a third-party model.
Once an answer reaches the assistant it is governed by the agreement between the user and that assistant's vendor. Cresva has no visibility into, and makes no representation about, how a vendor stores, processes or trains on content the user brings into their own conversation. A customer evaluating that risk should read the vendor's terms.
Cresva stores, for each connection: the registered client, the granting user, the brands and scopes granted, the time of grant, and the time the connection was last used. Access tokens and refresh tokens are stored as hashes rather than in a recoverable form.
Access tokens are short lived. Refresh tokens are single use: a refresh issues a new refresh token and invalidates the one presented, and presenting a token that has already been used revokes the grant. Records of a revoked grant are retained as audit history rather than deleted, so that past access remains attributable.
A user can revoke a connection at any time from Connected apps in their Cresva settings. Revocation takes effect immediately: the tokens are invalidated and the next call from that assistant is refused. Revoking one user's connection does not affect any other user's.
Removing a user from the Cresva account, or removing their access to a brand, has the same effect on what their connections can reach, because permissions are evaluated per call.
Cresva issues test API keys, distinguishable by their sk_test_ prefix, alongside live keys. A request authenticated with a test key reads simulated data only. It does not read the customer's connected accounts and cannot change any record. Test keys exist so that an integration can be built and run in a customer's own CI without touching production data.
Talk to us about compliance, data handling, or custom security requirements.
