Voice infrastructurefor enterprises

By default your data stays in the United States, hosted on Google Cloud with Cloudflare R2 storage, and inference runs from edge regions in the U.S. and Asia-Pacific (Tokyo) so your users get low latency wherever they are. For compliance-bound workloads, enterprise contracts can switch on Zero Data Retention, which means request text and audio are never written to disk. And if your data has to stay inside a specific country or region, the self-hosted enterprise tier runs fully inside your own infrastructure, so nothing ever leaves your environment.

Yes, and at serious volume. Capacity is provisioned as concurrent generations that scale with your contract, and we already have production customers running more than 1,000 concurrent generations. A Rust edge gateway serves inference across multiple GPU regions, so when your traffic surges our team can lift your limits the same day. You scale up without ever queuing behind a support ticket.

Security runs through every layer of the platform. Our SOC 2 Type II audit is currently underway, and the report will be available to customers under NDA once it is complete. Zero Data Retention is available on enterprise contracts, so request payloads are never persisted, and the self-hosted tier keeps every byte of your data inside your own environment. We also support HIPAA-aligned configurations and can sign a BAA for qualifying healthcare workloads, and independent penetration testing runs as part of our ongoing compliance program.

Absolutely. Enterprise customers get a direct line to our engineering team, not a ticketing queue, on whatever channel suits how your team works. We ship integration-specific features and protocol extensions for individual customers on a regular basis, and we stand up self-hosted deployments with you end to end, from first setup through go-live.

Yes, with fine-grained control from day one. Role-based access control lets you assign owner, admin, and member roles at the team level, plus manager, contributor, and viewer roles at the workspace level, so everyone has exactly the access they should. Single sign-on works today through Google and GitHub OAuth.

Both, and on your terms. You can spin up private voice clones from as little as 10 seconds of reference audio, 30 seconds or more for the best results, instantly through the API or the web UI, and they stay fully private to your team. For deeper engagements, we also fine-tune custom models on your own data.

Migrating to Fish Audio is straightforward, and most teams are surprised how quickly it goes. Your existing voices come across by recreating them from reference audio, our Python, TypeScript, and Go SDKs and WebSocket streaming API cover the integration patterns you already rely on, and our engineering team runs the cutover alongside you so production never skips a beat.