Open sourceMITopenwhispr.com← All comparisons

Platforms

Platform
InkSpoke
OpenWhispr
Windows
macOS
Apple Silicon + Intel.
Linux
Mobile
~iOS & Android on Pro
~Companion app a Pro perk; iOS "coming soon".

Feature by feature

Capability
InkSpoke
OpenWhispr
Works offline / on-device
Offline-firstLocal Whisper; nothing leaves the device unless you opt into cloud.
Fully offline with local modelsLocal mode transmits no audio; transcripts stay in local DB.
Speech engine
Whisper (local) + optional cloud ASR
Whisper + Parakeet (local), OpenAI/BYOK cloud
AI text refinement
IntegratedLocal LLM (LlamaSharp) or optional cloud, in one <600ms pipeline.
Cloud/local cleanup + voice-agent commands (GPT-5, Claude, Gemini...).
Tone presets
Built-in tone/format presets.
No tone/style preset feature documented.
Custom vocabulary
Managed custom vocabulary.
Custom Dictionary as ASR context hints; syncs across devices.
Per-app workspaces
First-class, user-configurable per-app workspaces.
Single global hotkey + agent, not app-specific profiles.
Wake word / hands-free
Hands-free wake word or hotkey.
Hotkey-only activation.
Types into any app
Any app + atomic clipboard restore
Global hotkey pastes at the cursor in any app.
Bring your own key (BYOK)
Bring your own OpenAI / Anthropic / local key.
OpenAI, Anthropic, Google, Groq, Tinfoil, AWS Bedrock.
Open source
~Open modelsRuns open-source ASR/LLM models; the app itself is proprietary.
MIT, 4.3k+ stars
Privacy model
On-device by default
~Local mode fully privateCloud mode is zero-retention with 30-day deletion caps.

The honest verdict

Choose InkSpoke if…

  • You want native, non-Electron performance with a <600ms latency budget.
  • You need a wake word, per-app workspaces, and tone presets.
  • You want a one-time perpetual license and atomic clipboard-restore guarantees.

Choose OpenWhispr if…

  • You want a completely free, forkable dictation tool and don’t mind Electron overhead.
  • You want built-in meeting transcription and the widest BYOK provider choice.
  • You’re a developer who wants to inspect or contribute to the source.

Where OpenWhispr is strong

  • Genuinely free and open-source (MIT) with an active, popular repo — fully forkable.
  • Unlimited local dictation costs nothing forever.
  • Broadest BYOK cloud LLM selection, plus a multi-step AI voice-agent mode.
  • Built-in meeting transcription with speaker labels and a notes system.

Where OpenWhispr falls short

  • No wake word — hotkey-only.
  • No per-app workspaces and no tone presets.
  • Electron-based — heavier footprint than a native app, no published latency numbers.
  • Cloud free-tier limits push regular cloud users toward $8/mo Pro.
Sources · verified July 2026

Compiled from publicly available information as of July 2026. Competitor names and trademarks belong to their respective owners. Details change often — check each vendor’s site for the latest.

You spoke. Now ink it.

Free with platform-hosted voice + text models. Pro unlocks BYOK, on-device models, and unlimited workspaces.