Molten: Local AI Chat
OnDemandWorld (Hong Kong) Limited · Productivity
Independent analytics for this App Store listing. AppTracker is not affiliated with OnDemandWorld (Hong Kong) Limited.
Chart standing · US
Screenshots
About
Molten brings AI to your devices—completely local, completely private. Run LLMs Offline Chat with models from Ollama, Swama (MLX-optimized for Apple Silicon), or built-in On Device Models. No cloud API, no tracking—your data stays your devices. Key Features Seamless self-hosted server switching and model selection, including on-device models. Real-time streaming responses with performance analytics (tokens/sec, eval rates). Native SwiftUI interface with markdown rendering, syntax highlighting, dark mode, and keyboard shortcuts. Floating panel for quick chats on macOS. Privacy Obsessed 100% local and private processing. Open-source for full audit on GitHub. No telemetry or cloud dependency. Download Molten and own your AI today.
What's new · 1.1
What's New in Molten 1.1 Accessibility & Polish • VoiceOver labels and improved hit targets throughout • Drag-to-dismiss keyboard on the message list (iOS) • Respects Reduce Motion accessibility setting • Dismissible error banner for connection issues • Clearer analytics labels ("input speed" vs "prompt eval rate") Settings & Connection • SecureField masking for bearer tokens and API keys • Clear connection status icons in Settings • TTS voices load once when Settings opens (no more polling) Stability • Fixed a data-loss bug where removing a model could delete conversations • Fixed database growing with duplicate model entries on every refresh • Fixed voice input not updating the UI during transcription • Fixed text-to-speech silently failing on iOS • Fixed hotkeys triggering system actions alongside custom shortcuts • Privacy hardening and analytics accuracy improvements • Numerous concurrency fixes for reliability Privacy: Molten collects no data. Ever.
Details