OpenWebUI vs MChatP

OpenWebUI Client is a native app for your own Open WebUI server — chat with your AI models with the fluidity of a real app, keeping everything on your server and under your control. FEATURES • Chat with your models, with real-time (streaming) responses • Multiple providers and models configured in your Open WebUI • Image generation • Notes and knowledge base (RAG) • Voice mode: speak and hear the responses • Chat actions: pin, archive, rename, clone, share, and export • Themes and customization • Attachments: send photos and files • 34 app languages, auto-detected from your device PRIVACY The app talks only to the Open WebUI server you specify. Nothing goes through third parties — your data stays on your server. REQUIREMENTS You need access to an Open WebUI server. On first launch, enter the server address and sign in with your account. This is an independent client, not affiliated with the Open WebUI project.

MChatP is a chat client for the AI server you run yourself. You choose the computer, the server software and the models: LM Studio, or any OpenAI-compatible server such as Ollama, oMLX or llama.cpp, with the open models you installed. MChatP hosts no model and runs no backend: it talks to your server, at home over your local network or from anywhere over a private network such as Tailscale, and shows the replies. Using MChatP requires a computer already running one of these servers (third-party software); the in-app Help walks through the setup step by step. YOUR SERVER, YOUR MODELS • Any model your server offers, from small and fast to large ones no phone could run, switched in one tap • LM Studio through its native API, with the MCP tools of your mcp.json: your files and tools stay on your computer, the server runs them • Ollama, oMLX, llama.cpp and any server exposing /v1/chat/completions, token optional • Several server profiles (MChatP Pro) MANAGE MODELS REMOTELY (MCHATP PRO) Load or unload a model on your LM Studio computer, or download a new one there from its catalog ID or a Hugging Face link, with live progress. PHOTOS AND DOCUMENTS • Show up to four photos to a vision model • Attach a PDF, text, Markdown, JSON or CSV file: its text is extracted on your device and sent with your message, so it works with any model WATCH IT THINK • The model's reasoning, folded above the answer • MCP tool calls as they start, succeed or fail • Model loading and prompt progress, then speed in tokens per second • Developer Mode adds time to first token, token counts and the full detail of each tool call (MChatP Pro) SYSTEM PROMPTS AND CHAT TOOLS A system prompt to steer the model (one included, a library with MChatP Pro); stop or regenerate a reply, copy, rename, search your history, keyboard shortcuts (⌘N, ⌘↩, ⌘K). HANDS-FREE • Ask MChatP from the action button or by voice; the answer comes back without opening the app • Start a conversation or test your server from the system's shortcuts PRIVATE BY DESIGN • Server profiles, tokens, conversations and attachments stay on your device • No account, no cloud sync, no analytics • Tokens in the device keychain; optional PIN lock with biometric unlock • Transparent about AI: you always know you are chatting with an AI model, the one you loaded on your own server, as the European AI Act requires MCHATP PRO, ONE-TIME PURCHASE Several server profiles, remote model management on LM Studio, a library of system prompts, Developer Mode and conversation export, unlocked with a single non-consumable purchase, charged at confirmation; the price is shown in the app. No subscription, nothing renews. After reinstalling or changing devices, Restore Purchases unlocks it again at no extra cost. MChatP is an independent app, not affiliated with the makers of LM Studio, Ollama, llama.cpp, oMLX or Tailscale, trademarks of their respective owners. Privacy Policy: https://precellence.icu/apps/mchatp/privacy_policy.html Terms of Use: https://precellence.icu/apps/mchatp/terms_of_use.html Support: support@precellence.icu