Your designs are protected by default. Full IP ownership, GDPR-compliant EU hosting, and workspace-scoped isolation, built into Carve from day one. The same rights apply to every plan, including Student.
Architectural drawings, sketches, and renders are high-value intellectual property. When AI tools enter the workflow, data governance cannot be an afterthought, it has to be a default.
Carve never uses your uploads, renders, or prompts to train models, on any plan. We only work with inference providers that contractually commit to the same.
All uploads and renders are stored encrypted at Wasabi EU in Frankfurt. During generation, images are sent to our AI providers only for as long as the render takes, then deleted.
Every prompt you write, from material descriptions to lighting tweaks, is your intellectual property. Stored alongside the render, isolated to your workspace.
The same ownership rules apply across every plan, Starter, Pro, Agency, Student, and Done-for-you. There is no enterprise-only tier that unlocks ownership.
Starter, Pro, Agency, Student, and Done-for-you. No enterprise-only tier unlocks ownership on Carve.
Carve treats every prompt and render as confidential information by default. There is no separate "paid plans only" version of this guarantee.
Carve never puts your uploads, renders, animations, or prompts into a training pipeline, and our inference providers (Replicate, kie.ai) commit not to train on API data either.
Carve never showcases your renders in marketing, on social, or in the public gallery without explicit written permission.
Your data only ever reaches the systems that actually process it, and only for the duration of the generation. Nothing flows to analytics brokers, advertising networks, or data buyers.
The infrastructure layer is set up to match the confidentiality promise above.
All uploads, renders, and animations are stored encrypted at Wasabi EU (eu-central-2, Frankfurt). During generation, images are sent to our AI providers for processing and deleted on their side afterward.
Every project lives inside a workspace tied to your organization. Only invited members can read it. Organization owners and admins inherit access, which is shown transparently in the members table.
All traffic is served over HTTPS. Uploads use short-lived presigned URLs scoped to your workspace, so even an intercepted URL cannot be reused.
Delete any render permanently from the dashboard. Delete an entire workspace from settings. We do not keep "soft-deleted" copies of your renders for marketing or research.
From competition entries to client presentations, your architectural data deserves the same confidentiality you apply to your practice.
Reliable AI infrastructure is essential when you're rendering to deadlines. Carve's operational policies keep performance consistent across every team.
Carve runs on shared infrastructure with each organization scoped to its own data envelope. Same security and performance for every team, with no resource starvation between accounts.
System health, queue depth, and credit usage are monitored so anomalies are caught before they impact your renders.
API endpoints and share-link tokens are rate-limited per IP and per token. Live-collab guests are capped per project so a leaked link can't fan out.
Every failed render, edit, animation, and upscale automatically refunds its credits to your workspace pool. You pay for results, not attempts.
Common questions from architecture firms, interior designers, and real estate teams evaluating Carve.
Carve never trains on your uploads, prompts, or renders. The inference providers we use (Replicate and kie.ai) also commit not to train on API data and delete it shortly after processing. This applies to every plan, including the Student tier.
Same ownership, same confidentiality, same EU hosting, on every plan.
