What is ShipAny Voice Agent?
ShipAny Voice Agent is conversational speech production built on the ShipAny SaaS foundation. Users generate speech, upload or record audio for transcription, and refine the result over several turns, with every version kept in a clear lineage.
An Audio Studio brings playback, transcript, download and version history together, and a media library stores every asset. ShipAny's auth, subscriptions, credits, payments, admin, storage and sharing are already wired.
Who it's for
- Founders launching an AI voice or TTS product
- Podcast, narration and dubbing tools
- Transcription services
- Developers building on OpenAI Speech or ElevenLabs
Key features
Text-to-speech
OpenAI Speech and ElevenLabs adapters for natural voices.
Multi-turn revisions
Ask for changes in conversation; each revision keeps an explicit link to the asset it came from.
Uploads, recording and transcription
Upload audio or video, record in the browser, and transcribe it.
Audio Studio
Playback, transcript, download and version history in one view.
Media library
First-class media asset records for everything users create or upload.
SaaS foundation
Auth, subscriptions, credits, payments, admin, storage and sharing from ShipAny.
What's included
- Foundation
- ShipAny (TanStack Start, React 19, TypeScript)
- Speech providers
- OpenAI Speech, ElevenLabs
- Agent LLM
- Configurable chat model in admin settings
- Storage
- Cloudflare R2 (required for generated speech)
- Uploads
- Audio and video up to 25 MB
- Deployment
- Cloudflare Workers or Node
Getting started
- 1
Install and run
pnpm install, copy .env.example to .env.development, run db:push and pnpm dev.
- 2
Configure providers
Set the chat LLM, OpenAI or ElevenLabs keys and R2 credentials in Admin → Settings.
- 3
Verify and deploy
Run the type check, tests and build, then deploy.
What you can build with ShipAny Voice Agent
AI voiceover tools
Narration for videos, courses and ads.
Transcription apps
Turn meetings, interviews or videos into text.
Podcast production
Draft, voice and revise episodes conversationally.
FAQ
Which TTS providers are supported?
OpenAI Speech and ElevenLabs, configured in the admin settings.
Is R2 required?
R2 is required for generated speech. Local uploads up to 25 MB work without it.
Can users revise generated audio?
Yes. Revisions happen in conversation, and the Audio Studio keeps the full version history.
Technical details (README)
ShipAny Voice Agent
Conversational speech production built on the ShipAny SaaS foundation.
The current MVP supports:
- Text-to-speech with OpenAI Speech and ElevenLabs adapters
- Multi-turn revisions with explicit asset lineage
- Audio and video uploads, browser recording, and transcription
- Audio Studio with playback, transcript, download, and version history
- Media library backed by first-class
media_assetrecords - Existing ShipAny auth, subscriptions, credits, payments, admin, storage, and sharing
Local development
pnpm install
cp .env.example .env.development
pnpm db:push
pnpm dev
Configure the chat LLM, OpenAI/ElevenLabs, and R2 credentials from Admin Settings. R2 is required for generated speech; local uploads up to 25 MB work without R2.
Verification
pnpm exec tsc --noEmit
pnpm test
pnpm build