Public beta — voice dictation, hosted in Australia

Voice typing that thinks.

Hold a key, speak in any language, paste polished text into any app. Zinga Flow transcribes, cleans up filler words and grammar, and lets you keep your flow.

faster than typing
100+
languages auto-detected
AU
data hosted in Australia
Power features

Beyond simple dictation.

Command Mode

Highlight any text and tell Zinga Flow what to do with it: translate, shorten, rewrite as a Slack message, convert to a bulleted list. Phase 1.

Transforms

Save reusable cleanup recipes — "polish for LinkedIn", "as a code review comment". Apply per dictation or set as default per app. Phase 1.

Scratchpad

A floating, sync'd dictation pad with tabs. Jot, refine, then send to your editor of choice. Phase 1.

Developer module

Cursor-, VS Code-, and terminal-aware. Picks up syntax cues, file names, and code context. Pre-trained on common stack jargon. Phase 1.

Core capabilities

Everything dictation should be — and nothing it shouldn't.

Universal dictation

Phase 1

Speak into any text field — browser, editor, terminal, chat. Coming on macOS first, then Windows, iOS, and Android.

4× faster than typing

Today

Average typing is ~45 WPM. Zinga Flow consistently lands above 200 WPM on the demo, including the AI cleanup pass.

100+ languages

Today

Auto-detected. No language switching, no manual settings. Works for English, Thai, Mandarin, Spanish, Vietnamese, Korean, more.

AI cleanup, your way

Today

Three levels — Light keeps your voice, Medium tidies grammar, High polishes for publication. You pick per dictation.

Custom dictionary

Phase 2

Teach Zinga Flow your industry jargon, names, and acronyms. Corrections compound the more you use it.

Snippets

Phase 2

Voice-trigger long pre-written text. "Email signature" → your full sign-off, instantly.

100+ languages

Speak naturally, in any language.

Zinga Flow auto-detects your language per dictation. Mix English and Thai mid-sentence — it keeps up. No settings to flip, no plugins to install.

EnglishภาษาไทยEspañolFrançaisDeutschPortuguês中文日本語한국어Tiếng ViệtBahasa IndonesiaBahasa Melayuالعربيةहिन्दीবাংলাItalianoNederlandsPolskiРусскийTürkçeУкраїнськаTagalogSvenskaNorskSuomiDanskภาษาลาวខ្មែរမြန်မာفارسی + many more
Privacy

Your audio never leaves Australia.

Most voice AI tools route your audio through US data centres. Zinga Flow runs on Australian infrastructure — Sydney-region GPU inference via our cloud-first.ai stack — so your speech stays under Australian privacy law and Aspire's own SOC 2 program.

  • Australian data residency by default — no opt-in required.
  • Privacy Mode option for zero-retention processing.
  • HIPAA-aligned processing for enterprise plans.
  • SSO and SCIM via Aspire Hub for teams.
Stack
Region
Sydney (AU)
ASR
Qwen3-Omni 30B
Cleanup
Qwen3.5 35B
Gateway
cloud-first.ai (LiteLLM)
Storage
MinIO (encrypted)
Retention
7 days · zero on Pro+
Teams

Made for collaborators.

Shared dictionary

One source of truth for product names, customer names, and internal jargon. Roll out corrections to the whole team instantly. Phase 4.

Shared snippets

Standard email replies, support boilerplate, on-brand answers — voice-triggered for everyone. Phase 4.

Aspire Hub SSO

Single sign-on, SCIM provisioning, role-based access — all via the existing Aspire Hub identity provider. Phase 4.

Compare

How Zinga Flow stacks up.

Honest about what we ship today vs. what's on the phased roadmap. Wispr Flow is the market-leading benchmark — we benchmark against them transparently.

FeatureZinga FlowWispr FlowSuperwhisperAqua VoiceDragon
AU-hosted inference
AI cleanup pass
macOS nativePhase 1
Windows nativePhase 1
iOS / AndroidPhase 3
Languages100+10450+~1040+
Custom dictionaryPhase 2
Snippets / triggersPhase 2
Team planPhase 4
Open pricing (AUD)$15/mo$22/mo$12/mo$8/mo$300+

Source: vendor websites + Wispr Flow research report (May 2026). Pricing approximate, USD converted to AUD at 1.52.

Stop typing. Start flowing.

Try the in-browser demo right now — 60 seconds, no signup required.