About Private Chat Hub

Privacy-first AI on your terms. Own your conversations.

About Private Chat Hub

Private Chat Hub is a privacy-first Android app for chatting with AI models you run yourself—on your local network or entirely on-device. Connect to Ollama, LM Studio, or OpenCode, or run Gemma / Gemini Nano with zero internet access. Your conversations never leave your control.

🎨 Our Philosophy

Privacy First: User data never leaves your control—no telemetry, no cloud sync

Local Performance: Optimized for local network and on-device operations

Simplicity: Complex AI capabilities made accessible through a clean interface

Transparency: Always show what's happening—active model, token counts, tool calls

Extensibility: Multiple backends, pluggable tools, and open-source codebase

👥 Who Is This For?

The Privacy Advocate

Runs home servers and self-hosted services. Distrusts cloud AI. Values complete data sovereignty. Wants an AI assistant that never phones home.

The AI Developer

Experiments with LLMs, tool calling, and multi-model workflows. Needs a flexible testing client that exposes real API behaviour—streaming, function calls, thinking traces—without rate limits or costs.

The Power User

Uses AI daily for writing, coding, and research. Values TTS read-aloud, model comparison, project workspaces, web search integration, and offline capability on a single mobile app.

The Offline User

Needs AI in environments without internet. Uses on-device Gemma/Gemini Nano inference that runs entirely on the Android device—no server, no network.

🤖 Supported Backends

  • Ollama — Self-hosted LLM server. Supports the full Ollama model library with tool calling, vision, and thinking modes. Default port 11434.
  • LM Studio — Run any GGUF/GGML model from LM Studio's desktop app over your local network. Default port 1234.
  • OpenCode — Routes requests to cloud providers (OpenAI, Anthropic, Google) via the self-hosted opencode server. Default port 8080.
  • On-Device (LiteRT) — Gemma 3 / Gemini Nano running natively on the Android device via Google's LiteRT runtime. Completely offline, no server required.

🏗️ Built With

  • Flutter 3.10.1+ — Cross-platform Android development
  • Dart 3.10.1+ — Fast, type-safe language
  • Material Design 3 — Beautiful dark UI with adaptive theming
  • LiteRT (TensorFlow Lite) — On-device model inference
  • flutter_markdown + LaTeX — Rich formatted responses with math
  • Jina AI — Web search with source links
  • flutter_tts — Text-to-speech read-aloud

🎯 Product Principles

  1. Data Ownership: Users own 100% of their conversation data
  2. No Subscriptions: Free to use; costs only your own hardware
  3. Backend Flexibility: Swap between Ollama, LM Studio, OpenCode, or on-device at any time
  4. Model Flexibility: Switch models instantly within any conversation
  5. Offline First: On-device inference works with zero network
  6. Open Development: MIT-licensed, built in the open

🔗 Project Links

GitHub Repository • Releases & Downloads • Installation Guide • All Features • API Comparison