Security
How your data is protected, what you can take with you, and what we’re building next. Statuses are labeled; nothing here is aspirational unless it says so.
The stance
Your contracts, your standards, and your team’s decisions are some of the most sensitive commercial information your company holds. Two commitments sit under everything else on this page: your data is never used to train AI models, and everything you build in the platform is an asset you can take with you. A tool that holds your negotiating history has no business holding it hostage.
What runs today
These are the controls in the shipped platform, not a roadmap:
Workspace isolation
Isolation is architectural. Every request is scoped to a workspace, and your organization’s data does not cross customer boundaries.
Role-based access
Access is controlled at two levels: what someone may do to the account, and what they may do inside each workspace. Reaching a workspace requires an explicit membership.
The audit log
Actions on your contracts, evaluations, and decisions are recorded with who did it and when. Decisions carry attribution, so the record of a negotiation shows whose judgment it reflects.
Encryption in transit
All traffic to the platform is encrypted in transit over TLS.
US processing
The platform is operated from the United States and its territories, and data is processed there.
Human-gated output
Nothing reaches your vendor without a human decision. Exports are blocked while designated must-address findings remain undecided.
AI processing
Contract evaluation is powered by frontier AI models accessed under commercial terms that prohibit training on your data, and retrieval runs on embedding models under the same kind of no-training terms. The specific providers can change as the frontier moves, which is exactly why we don’t carve names into a marketing page: the current subprocessor list, with names, is available to any customer or security reviewer on request.
What the platform learns from your team’s decisions is stored as human-readable records inside your workspace, not trained into anyone’s model weights. That distinction is load-bearing: it is why your institutional knowledge stays yours, survives any change of underlying model, and leaves with you if you go.
Your data leaves with you
The platform’s value compounds the longer you use it, and that is exactly why the exit has to be honest. Everything your organization accumulates is exportable in full:
- Your contracts, as the files you uploaded
- Every evaluation, finding, and coverage receipt
- Your standards, as readable prose your team can use anywhere
- Your decision history: what your team decided, on what language, and why
Export is available throughout your agreement, not just at the end, and it is walked through with you before any agreement ends. If Revver disappeared tomorrow, your institutional knowledge would not.
Deployment options
Standard deployments run on managed cloud infrastructure in the United States, and for most companies that is the right answer: running in days, nothing to install, nothing to integrate before the first contract is read.
For organizations whose security posture requires more, dedicated single-tenant deployment is available on enterprise engagements, including AWS-hosted environments with model processing through Amazon Bedrock, so contract text never leaves infrastructure scoped to you. We build that deployment when your requirements call for it, scoped with your security team, rather than maintaining a one-size demo of it. If your review needs this tier, say so in the first conversation and it becomes part of the engagement.
The roadmap, plainly labeled
Revver is an early company and this page does not pretend otherwise. Here is what is in motion, with honest statuses:
Independent security assessment: scheduled
An outside security firm is engaged to run a full assessment of the platform. Findings drive the hardening queue before anything else does.
SOC 2: on the roadmap
SOC 2 attestation is the destination; the independent assessment above is the first step on that path. We will publish status here as it progresses.
Enterprise controls: built on engagement
SSO/SAML, custom data-processing agreements, and dedicated deployment are scoped per enterprise engagement rather than shipped speculatively.
Ask us anything
Running a vendor-security review? Send the questionnaire to sp@revver.ai. You will get straight answers, including “not yet” where that is the true answer, because a security review is exactly the wrong place to discover a vendor who rounds up.