An AI voice companion with an animated character, hands-free wake mode, bring-your-own-key AI providers, ElevenLabs speech, optional vision, and saved past chats.

Capabilities
BYOK AI, speech, voice, hands-free wake, camera context, and local conversation history in one iOS app.
Speak from the iOS composer using native iOS speech or your configured cloud STT provider.
When enabled, Scowld can listen locally for Bella or your custom name to start voice recording.
A compact in-app guide explains wake phrases and the eye, hands-free, voice, send, and cancel controls.
A VRM character with lip sync, idle animation, and expressive response playback.
Enable camera context when you want the AI to respond to what you are seeing.
Save conversations locally, switch between chats, and reuse one as context for future replies.
Use your own AI, speech-to-text, and ElevenLabs keys. Keys stay in the iOS Keychain.
Celine, Claire, bundled voice presets, custom voice ID support, and local voice previews.
FAQ
Free, open source, and bring-your-own-key. Here is how Scowld works.
Yes. Scowld is free and open source (MIT). There are no subscriptions, paywalls, voice credits, or in-app purchases. You bring your own provider API keys and pay your chosen providers directly for usage.
You use your own API keys for the AI, speech-to-text, and text-to-speech providers you choose. Keys are stored only in the iOS Keychain on your device, so Scowld never runs a server that sees or stores them.
Gemini, OpenAI, Claude, Ollama, Groq, OpenRouter, xAI, Together AI, Hugging Face, Venice AI, and Moonshot AI. You can switch providers and models any time in Settings.
Voice input works with native iOS speech recognition or cloud providers like OpenAI Whisper, Groq Whisper, Deepgram, AssemblyAI, and Google Cloud Speech-to-Text. Spoken replies use ElevenLabs or OpenAI voices.
Yes. Provider keys are stored in the iOS Keychain on your device, are not bundled in the app binary, and are never sent to Scowld servers. Requests go directly from your device to the provider you configured.
Yes. Scowld is released under the MIT License and the source is available on GitHub at github.com/apoorvdarshan/scowld.
No account is required. Cloud AI, speech-to-text, and text-to-speech need an internet connection and valid provider keys, while native iOS speech recognition works on device.
An optional mode where Scowld listens on device for "Bella" or your custom companion name to start voice recording. Wake detection is processed on device and is not sent to providers before recording starts.