Local-first / Unlimited dictation
Your voice.
Direct to cursor.
Local dictation is unlimited and free. In local mode, your voice never leaves your PC.Behavior: desktop/electron/main.ts local sidecar path
Moonshine-tiny warm median 0.15s on the repository's CPU benchmark fixture.Measurement: reports/bench_results_20260710_041933.json
Pay only for the intelligence on top: cleanup and SeamlessM4T translation across 85 supported languages.Behavior: desktop/src/languages.ts language catalog
Hold Ctrl+Winand just talk
English transcript
Turn every spoken thought into text, right where you work.
German translation
Verwandle jeden gesprochenen Gedanken in Text, genau dort, wo du arbeitest.
01 / Product journey
From rough speech to ready text
Each layer solves the next interruption between a thought and the text you meant to write.
- 01
Custom Vocabulary
ProblemNames, brands, and team jargon arrive misspelled.
OutcomeYour saved vocabulary repairs exact terms before they reach the cursor.
- 02
Local AI Cleanup
ProblemGood ideas still contain filler, false starts, and rough punctuation.
OutcomeOptional on-device cleanup polishes the finished transcript, with the original preserved if cleanup fails.
- 03
Translation
ProblemThe person reading works in a different language.
OutcomeA second hotkey uses SeamlessM4T to dictate and paste across 85 supported languages.Behavior: desktop/src/languages.ts language catalog
- 04
Developer Mode
ProblemSpoken code loses its casing and symbols.
OutcomeDeveloper formatting turns spoken conventions into code-ready text.
- 05
Per-App Profiles
ProblemOne writing style does not fit every window.
OutcomeThe foreground app selects its own tone and processing steps automatically.
- 06
Voice Commands
ProblemSmall edits pull your hands back to the keyboard.
OutcomeA whole-utterance command performs the edit and pastes no stray text.
- 07
Snippets
ProblemSignatures, addresses, and boilerplate repeat all day.
OutcomeA spoken trigger expands into the exact saved text in one breath.
02 / Privacy
Local mode vs cloud mode
Wispr Flow is a subscription alternative. Here is how SpeechWalker's local and cloud paths differ on the things that matter.
| SpeechWalker | SpeechWalker Cloud | |
|---|---|---|
| Privacy | In local mode, your voice never leaves your PC Behavior: desktop/electron/main.ts local sidecar path | Cloud mode uploads audio to transcribe Behavior: desktop/electron/cloudTranscribe.ts |
| Offline use | Local dictation works offline after the model is downloaded Behavior: cache-first loading in desktop/python/transcribe_service.py | Cloud mode requires an internet connection Behavior: desktop/electron/cloudTranscribe.ts |
| Price | Dictation free forever, €49 once for the intelligence | $144/yr, forever — $432 over three years |
03 / Requirements
Choose the mode that fits your PC.
Cloud transcription is optional. Local dictation remains available where supported, and no automatic overage purchase occurs.
| Requirement | Fast local transcription | Local translation | Cloud mode |
|---|---|---|---|
| OS | Windows 11 x64 | Windows 11 x64 | Windows 11 x64 |
| Memory | 8 GB RAM | 16 GB RAM recommended | 8 GB RAM |
| Compute | CPU-only ready | Capable GPU recommended | Light local compute |
| Connectivity | Offline after model download | Offline after model download | Internet required |
| Privacy | Audio stays on your device | Audio stays on your device | Audio is sent to the cloud provider |
| Expected speed | Designed for fast dictation | Expected to vary by model and hardware; not a measured benchmark | Depends on connection and provider |
04 / Pricing
Honest Pricing
Local dictation is free forever. Pay once to unlock the intelligence on top.
Pro Lifetime
- Own it forever — no renewal or cancellation
- On-device translation (85 supported languages)Behavior: desktop/src/languages.ts
- All local models unlocked
- Local-mode audio stays on your deviceBehavior: local sidecar path
Cloud
or €49/yr — best value
- Cloud transcription for any hardware
- No fixed minute allowance
- Renews monthly or yearly until cancelled
- Cancel anytime
- Local dictation remains available as fallback
- Stacks on Free or Pro
Team
or €120 per seat/year · 3 seats minimum
- 300 Included Minutes per seat every billing cycle
- Renews monthly or yearly until cancelled
- Local fallback when cloud credit is unavailable
- Shared vocabulary across the team
- Seats you manage, each carrying its own Pro entitlement
- Central billing
Wispr Flow costs $144/yr — $432 over three years.
SpeechWalker Pro is €49 once. Own it forever.
05 / FAQ
Frequently asked questions
- Is SpeechWalker private, and how does it protect privacy?
- Yes. In local mode, audio stays on your device. Audio is uploaded only when you choose cloud transcription.
- Does SpeechWalker work offline?
- Local dictation and local translation work offline after the selected model is downloaded.
- What hardware does local dictation need?
- SpeechWalker requires Windows 11 x64 and 8 GB RAM. Local transcription works on a CPU; local translation is best with 16 GB RAM and a capable GPU.
- Which languages can SpeechWalker translate?
- On-device translation supports the 85 supported languages listed by the app.
- Do I need cloud transcription?
- No. Cloud transcription is optional and intended for hardware that cannot run the local models comfortably.
- How much does SpeechWalker cost?
- Local dictation is free. Pro costs €49 one-time, optional Cloud costs €4.99 per month, and Team costs €12 per seat per month with a three-seat minimum.
- Do I need an account?
- No account is required for free local dictation. An account is required to buy and manage paid plans.
- What is SpeechWalker Team?
- Team adds shared vocabulary, central billing, managed seats, and 300 cloud minutes per seat each billing cycle.
06 / RELEASES
Get the Windows release notice
If the verified installer is not available in your account yet, leave your email and we will notify you when it is ready.