Capabilities

Everything you need. Nothing you didn’t ask for.

Built for voice, optimised for mobile, works in your browser. Free forever.

Get started free →See pricing

Native voice-to-text

Uses your device OS speech engine. No third-party voice API, no audio transmitted — transcription runs entirely on-device.

Zero voice retention

We receive text, never audio. No voice profile is created. Nothing to hack, sell, or leak.

Install to home screen

A PWA you add to iPhone or Android in one tap from your browser. No App Store. Full-screen, offline-ready.

Multiple AI models

Auto mode picks the fastest available model. Groq Llama is always free. Gemini Flash Lite for smarter replies.

AI image generation

Say "create an image of…" and Friday generates it instantly. Unlimited basic images free. Pro quality via image credits.

Free forever tier

100 messages per day and unlimited basic images, free forever. No hidden limits, no trial expiry, no surprises.

Persistent chat history

Friday remembers your conversations across sessions and devices. Free: 7 days. Premium: 30 days. Premium+: 90 days.

Secure sign-in

Google OAuth, email + password, or magic link. Row-level security ensures your data is only ever accessible to you.

"Hey Friday" wake word

Hands-free activation on Android/Chrome. On iOS, create a Siri Shortcut to open Friday and start listening immediately.

In depth

How the parts that matter actually work

Voice that stays on your device

Every time you speak, your browser or phone OS handles speech-to-text locally. Friday receives only the transcribed text — no audio waveform, no voice sample, no recording.

This works via the Web Speech API (Chrome, Edge, Safari) and native OS dictation engines on iOS and Android. You can choose from any voice installed on your device for replies, including Microsoft Neural and Apple Neural voices.

  • No audio sent to any server
  • No voice profile created
  • Works offline for speech capture
  • Premium Apple/Microsoft neural voices

Smart model routing — automatic or manual

In Auto mode, Friday picks the best available model for each message based on your plan and current API quota. On the free tier, Groq Llama 3.1 is always available with no daily limit.

Manually switch between models from the model selector in the chat header. Each model shows its credit cost before you use it.

  • Llama 3.1 — always free, instant
  • Gemini Flash Lite — 1 credit, smarter
  • Gemini 3.5 Flash — 2 credits, faster reasoning
  • Gemini 3.5 Pro — coming soon on Premium+

AI image generation built in

Just ask naturally — "create a birthday card for my mum with flowers" — and Friday generates an image and shows it inline in the conversation.

Basic images use Pollinations FLUX and are completely free with no credits required. Pro-quality images use Imagen and require image credits, earned by watching optional rewarded ads or purchased in packs.

  • Unlimited basic images, free forever
  • Earn credits via optional ads
  • Buy credit packs from $1
  • Credits never expire

Getting started

Up and running in under a minute.

  1. 01

    Open the app

    Visit app.heyfriday.co in Safari or Chrome. Sign up with Google in under 10 seconds. No credit card required.

  2. 02

    Speak or type

    Hold the mic button and speak naturally. Your OS transcribes it on-device. Friday receives only the text.

  3. 03

    Get an answer

    Intelligent responses delivered instantly. Ask for code, summaries, images, plans, or anything else.

Try Hey Friday free

Voice-first AI assistant that runs in your browser. No download, no credit card.

Get early accessSee what it does