← Vendor Profiles
Voice inputMedium sensitivity

Wispr Flow

An AI voice-dictation app used for fast voice-to-text input across every desktop and mobile app, turning speech into cleaned, formatted writing in any text field.

wispr.ai
Use case

AI-powered voice-to-text dictation and writing assistant.

Access controls

2FA · local-only audio

Swap in

Superwhisper · MacWhisper · Aqua Voice

Vendor profile
What it is

An AI voice-dictation application ("Voice OS") from Wispr AI that turns natural speech into polished, formatted text across virtually any app on desktop and mobile.

Core functionality

A hold-to-talk shortcut works in any text field across macOS, Windows, iOS, and Android. Audio is transcribed in the cloud and AI-cleaned (removes filler words, auto punctuates/capitalizes, structures lists, adapts tone per app). Command Mode (Pro) lets users voice-edit highlighted text ("make this concise," "translate to Polish"). Supports 100+ languages with code-switching.

AI & data capabilities

Core is its speech-to-text + AI cleanup pipeline (reportedly using an OpenAI subprocessor for transcription plus a fine-tuned Llama model for cleanup, hosted on AWS via Baseten). Claims ~10% error rate (vs. higher for generic dictation). Building proprietary voice-first foundation models. No MCP server; this is an input tool, not an integration platform.

Pricing

Per-seat subscription. A free tier plus paid plans roughly in the $12–$15/user/month range (Pro). Enterprise plans available with additional security (SOC 2 Type II, auto-enabled privacy mode).

Integrations & ecosystem

Works system-wide across apps (not via discrete integrations); a closed API has been tested with select enterprises/hardware partners.

Security & compliance

SOC 2 Type II (Enterprise; 2026 re-attestation in progress via Drata/A-LIGN), ISO 27001:2022, HIPAA BAA available on all plans, GDPR/CCPA aligned, ISO 42001 (AI governance) alignment claimed. Privacy Mode (zero data retention) is opt-in for Pro, auto-enabled for Enterprise; without it, dictation data "may be used to evaluate, train and improve" the product.

Company background

Wispr AI Inc., founded 2021 in San Francisco by Tanay Kothari (CEO) and Sahaj Garg; originally a wearable hardware company before pivoting to software. Flow launched on macOS October 2024. Raised $81M total ($30M Series A led by Menlo Ventures, June 2025; $25M extension led by Notable Capital, November 2025) at a $700M post-money valuation. Bloomberg reported on May 12, 2026 that Wispr was in talks to raise ~$260M led by Menlo at a ~$2B valuation (not finalized). ~2.5M downloads; used at 270 of the Fortune 500.

VC / GE fit

Productivity tool for individuals — faster drafting of emails, memos, notes, and messages by voice (claimed ~4x faster than typing). Useful for investors who write heavily. Not a data/sourcing tool.

Limitations

Cloud-dependent (offline tools run locally; vendor outage/acquisition risk); AI cleanup can over-edit/alter verbatim meaning; reported high RAM/CPU usage on the Windows (Electron) app; for sensitive content, privacy mode must be enabled (default training use otherwise); compliance re-attestation reportedly in progress.