isper

Simple, Transparent Pricing

Xisper is free to use. Want no limits? Run it yourself — the whole stack is open source.

Free

$0/ forever

For casual users and trying out Xisper

Download Free
75 minutes of speech recognition per week
10,000 characters per week
900 AI processing calls per day
Role-based vocabulary (Developer, Lawyer, Doctor, PM)
Custom hotwords and corrections
Local audio storage
Cross-device sync
Open Source

Self-Hosted

$0/ bring your own keys

For developers who want no limits and full control

View on GitHub
No quotas — you pay your ASR and LLM providers directly
Full source: macOS app, Workers backend, AI gateway
Runs on your own Cloudflare account
Role-based vocabulary (Developer, Lawyer, Doctor, PM)
Custom hotwords and corrections
MIT licensed

The hosted service is free within quota — no credit card, no checkout. Need more? The entire stack is open source; self-host it and use your own provider keys.

Frequently Asked Questions

What happens when I reach my quota?

When you reach your weekly speech recognition limit or daily AI processing limit, the app will notify you. You can wait for the quota to reset (weekly on Monday at 03:00 Beijing Time for ASR, daily at 03:00 Beijing Time for AI processing), or self-host Xisper with your own API keys for unlimited use.

Do I need to pay for anything?

No. Xisper is free to use within the quotas above, and there is nothing to buy. If you need more than the free tier, the entire stack is open source — self-host it with your own ASR and LLM keys and there are no limits.

How is speech recognition time calculated?

Speech recognition time is measured by the actual duration of your audio recordings. For example, if you record a 5-minute meeting, that counts as 5 minutes towards your weekly quota.

What are AI processing calls?

AI processing calls are used when you apply post-processing features like summarization, formatting, or other AI-powered enhancements to your transcribed text. Each operation counts as one call.

Is my data stored on your servers?

No. Your audio recordings and transcription text are stored locally on your device. We only store your account information, usage metrics, and custom vocabulary settings (hotwords and corrections) for cross-device sync.

What do I need to self-host?

A Cloudflare account (Workers, KV, R2), a Logto tenant for auth, and API keys from the ASR and LLM providers you want to use. The repository README walks through the full setup.

Ready to get started?

Download Xisper for free and experience role-based speech recognition.

Download Free