Private Chat Hub

🎯 The Problem

Cloud AI services lock you into their infrastructure and send your conversations to corporate servers. Private Chat Hub is different—all your conversations stay on your devices and infrastructure, whether you're using a local server or a completely offline on-device model.

✨ Why Private Chat Hub?

Feature	Private Chat Hub	ChatGPT / Claude
Data Ownership	✅ 100% User Controlled	❌ Cloud Storage
Model Choice	✅ Any Model (Ollama, LM Studio, On-Device)	❌ Provider Locked
Works Fully Offline	✅ On-Device Models	❌ Internet Required
Tool / Function Calling	✅ Built-in (Web Search, Location, URL Fetch…)	⚠️ Platform Dependent
Vision / Image Input	✅ Supported	✅ Supported
Text-to-Speech	✅ Built-in TTS	❌ Not Included
Cost	✅ Free (self-hosted)	💰 Subscription
Open Source	✅ MIT License	❌ Proprietary

🤖 Supported AI Backends

Connect to the backend that fits your setup—or combine several at once:

• Ollama — Any model from the Ollama library (Llama 3, Mistral, Gemma, Qwen, DeepSeek, and more). Tool calling fully supported.
• LM Studio — Run any GGUF model locally via LM Studio's desktop app. Connect over your local network.
• OpenCode — Route requests through the OpenCode server to cloud providers (OpenAI, Anthropic, Google). Tool calling handled by the provider.
• On-Device (LiteRT / Gemma) — Completely offline inference using Gemma or Gemini Nano running directly on your Android device. No server needed.

🚀 Core Features

• Real-Time Streaming — Responses stream token by token as the model generates them
• Tool / Function Calling — AI can call built-in tools: web search, current time, device location, URL fetch, and notifications
• Web Search Integration — Powered by Jina AI; results include clickable source links
• Vision & Image Input — Attach photos or files; vision-capable models can see and describe them
• Text-to-Speech — Have any AI response read aloud with built-in TTS streaming
• Model Comparison — Run the same prompt side-by-side across two or more models
• System Prompt per Conversation — Set a custom persona or instructions for each chat
• Project Workspaces — Organize conversations into named projects
• Offline Message Queue — Messages sent while offline are queued and delivered when reconnected
• Conversation Export — Export chats as JSON or Markdown; you own your history
• Markdown + LaTeX Rendering — Rich formatted responses with math support
• Thinking / Extended Reasoning — Support for reasoning-focused models with visible thought traces
• Multiple Model Switching — Swap models instantly within any conversation
• Material Design 3 — Beautiful dark UI with adaptive theming

👥 Built For

Privacy Advocates — Run your own infrastructure; no data leaves your network

AI Developers — Test and compare models rapidly, experiment with tool calling, inspect streaming responses

Power Users — Advanced features: tool toggles, model comparison, project workspaces, TTS, LaTeX

Offline Users — Fully on-device inference with Gemma/Gemini Nano, no internet required

⚡ Get Started

Download the APK, connect to your preferred backend, and start chatting. No cloud sign-up required.

📥 Download Latest Release →

📚 Learn More

All Features • About the Project • Installation Guide • API Comparison • GitHub