EN → ZH
A 28-minute tech interview re-voiced in Mandarin with the host's cloned voice; script reviewed line-by-line.
Podcast → multilingual, dubbed Your podcast, in every language
Your podcast, in every language
without losing your voice.
Drop your episode in. IceFold transcribes, segments, translates, and re-voices it across the languages you pick. The voice can be cloned, the script can be edited, every step is reviewable.
- High-quality transcripts with timestamps (Whisper + stable-ts)
- Voice cloning per language (ElevenLabs / MiniMax BYO key)
- Edit the translated script before voicing — no black-box outputs
- Re-run a single language without re-transcribing
Free to try. Bring your own provider key (OpenRouter / OpenAI / ElevenLabs / MiniMax) — IceFold doesn't markup AI tokens. Cancel anytime.
3-step pipeline
How it works
01
Drop your episode
Audio or audio-from-video. IceFold transcribes and splits into utterances with stable timestamps you can correct.
02
Pick target languages
Configure a language variant axis. The translation and voice-synth nodes fan out automatically across each target.
03
Approve script, re-voice, publish
Edit any translated line, pick a voice per language, generate audio. Re-running a fix is one click — IceFold reuses everything upstream.
Output examples
What you get
Real outputs from this flow. Every variant is fully editable — re-run a single language or platform without redoing the whole pipeline.
EN → JA
Japanese dub with native-pace pacing, automatically re-timed so the dub matches the original cadence.
EN → ES
Spanish dub with two voice options (warm female / deep male) generated as parallel variants.
Examples shown are produced by the Podcast Studio template. Open the studio to see the full DAG and tweak any node.
Pricing
Free to try. $5/month when you want to share.
Free
$0
- All built-in nodes & studios
- Multi-variant canvas
- 20 free runs / lifetime
- BYO provider key → unlimited
Most popular
Pro
$5/mo
- Everything in Free
- Install private Python nodes
- Publish flows to the Market
- Unlimited platform-hosted runs
FAQ
Common questions
Can I correct the transcript?
Yes. Every transcript line is editable before it flows into translation. If you fix the source, only the affected downstream variants are re-run.
Which TTS providers are supported?
ElevenLabs and MiniMax for voice synthesis, plus any OpenRouter audio model. Bring your own API key on the Pro plan.
Does IceFold do voice cloning?
IceFold wires up the provider you choose — voice cloning quality is whatever ElevenLabs / MiniMax give you with your key. You upload a reference, configure the voice in the node, run.
What about RSS / publishing?
Out of scope for IceFold — we hand you the final audio files. Plug them into your existing podcast host.
Episode length limits?
Free trial uploads are capped at 30 seconds for video and a few minutes for audio. Pro removes the cap.
Ready to run the Podcast Studio?
Open the template, swap your inputs, hit run. The first 20 runs are free — no credit card required.