Chat for Ollama vs opencode companion

Chat for Ollama is a client for the Ollama server you already run yourself — on a Mac under the desk, in a homelab, or across your own private network. Your questions go to your machine and nowhere else. The app needs a running server: it does not run models on the phone and does not download any. EVERYTHING YOUR SERVER CAN DO IS FREE - Streaming answers from every model on your server - Switch models mid-chat, the conversation stays - Thinking output from reasoning models, collapsed above the answer - Image input for models with vision - System prompt, temperature, Top-P and context length per chat - Unlimited history, stored on the device No server feature is behind a purchase. What you pay for is convenience the app adds, not something Ollama already gives you. CONNECTING, WITHOUT THE GUESSWORK Most first attempts fail for one reason: Ollama listens on 127.0.0.1 by default, so the phone cannot reach it — and no error message says so. - Scan your local network for servers on port 11434 - Paste a copied address straight from the clipboard - When a connection fails you get a screen, not an error code: what was tried, what came back, what it means, and the exact command to fix it with a copy button - Reachability dot per server THE CHAT ITSELF - Markdown, tables and code blocks with a copy button per block - Speed readout after each answer: tokens per second and duration - Context meter that warns before the server silently truncates the history - Copy, regenerate, edit your own message, delete - Stop button that really cancels the generation and keeps the partial text - Loading indicator with a counter when a cold model has to be read from disk - Chat list with automatic titles, search and pinning - Split-view layout on iPad ALSO WORKS WITH OPENAI-COMPATIBLE SERVERS Anything serving /v1/chat/completions on a machine you run — LM Studio, llama.cpp, vLLM, or Ollama's own compatibility endpoint. Same chat, same streaming, same cancelling. ONE-TIME UNLOCK, $5.99 No subscription and no second tier. The unlock adds: - Several servers with quick switching - Named presets for system prompt and parameters - Attachments: text, Markdown, code and PDF files are extracted as text - Export a chat as Markdown PRIVACY No account, no sign-up, no telemetry, no analytics, no crash reporting. Chats and server addresses stay on the device, API keys go into the Keychain. A network capture shows nothing but the server you entered yourself. WHAT YOU NEED An Ollama server (or an OpenAI-compatible one) that your iPhone can reach — same network, or through a mesh VPN you already use. There is a demo mode if you want to look around first. Chat for Ollama is an independent app. It is not affiliated with, endorsed by or connected to the Ollama project.

opencode companion puts the opencode coding agent in your pocket. Run opencode on your computer, then follow the work from your phone: watch replies stream in, approve permissions, review diffs and send the next instruction from anywhere. Your code never leaves your machine. The app connects to your own opencode server — over your local network, Tailscale or an end-to-end encrypted relay — and no account is needed. WHAT YOU CAN DO • Follow sessions live: streaming replies, reasoning and tool calls • Approve or reject tool permissions from your phone • Answer the agent's questions without leaving the app • Review every diff the agent produced, file by file • Browse and read your project files • Compose with @ files, / commands, ! shell and photo attachments • Keep several servers (LAN, Tailscale, relay) with automatic reconnect • Get the opencode terminal aesthetic on a small screen SETUP On your computer, run: npx ocremote --pair A QR appears; scan it once and you are connected. Keep the terminal running while you work. No computer at hand? Tap “Try the demo” to explore the app with sample data — nothing is sent anywhere. opencode is an open source AI coding agent by its authors; this app is an independent client for it.