LLM Chat

**LLM-Chat is a powerful, free chat client for Ollama that gives you AI assistance on your own terms.** Connect to your local server on the network or to Ollama Cloud using your own API key. No purchases, no subscriptions, no vendor lock-in – your chats stay on your device. ### ROCK-SOLID CHAT • Real-time streaming responses with speed (tokens/second) • Conversation history with folders, tags, bookmarks, and pinned messages • Fork a chat from any message to explore alternative paths • Change the theme per chat – 12 built-in themes including Dracula, Synthwave, Matrix, Nord, Catppuccin, Gruvbox, and Rosé Pine • Smart chat titles (AI-generated) and suggested follow-up questions ### POWER-USER TOOLS • 40+ slash commands: /python, /js, /sql, /json, /regex, /qr, /remind, and more • Run Python code and JavaScript directly on the device • In-memory SQLite playground with tabular output • RAG: index PDF and text documents and ask questions using semantic search • Research mode: automatic web searches via DuckDuckGo (no API key required) with summaries and sources • Tool mode: the model can use the app's tools (time, search, chat statistics) • Best of 3: generate three responses and let the AI choose the best one • A/B testing: compare two models side by side • Create custom plugin commands with prompt templates ### AGENT & MEMORY • Agent mode for action-oriented responses • Chat memory that summarizes and preserves important facts per chat • Automatic AI-powered tagging of responses ### IMAGES • Image generation via ComfyUI with FLUX (supports custom workflow files) • Image generation via AUTOMATIC1111 with any SD model • Send images to multimodal models, with camera and gallery support ### VOICE • Dictation with automatic stop on silence and auto-send • Push-to-talk • Voice commands (e.g. “summarize latest”, “research …”) • Text-to-speech in any language with speed control ### STATISTICS & CONTROL • Dashboard with token usage, speed, and generations per model • Full control: temperature, top-p, top-k, Mirostat, repeat penalty, seed, context size, and auto-trim • Export chats as Markdown, PDF, JSON, HTML, TXT, or CSV • Backup & restore + automatic daily backups ### MORE • Apple Watch app that connects directly to your server • iPad support with keyboard shortcuts (⌘N new chat, ⌘F search, ⌘, settings) • Personas: create your own system instructions with categories **NOTE:** LLM-Chat is a client. You need an Ollama server (local on your network or Ollama Cloud) and your own models. Everything in the app is