Sesori vs Tailscode

Sesori is the remote control for the AI coding agents running on your own computer. Start a task at your desk, then follow it, answer it and steer it from your iPhone. Your code stays on your machine. Works with OpenCode, Codex, GitHub Copilot, Cursor, Claude Code, Pi, OMP, Hermes, DeepSeek and Grok Build. Sesori connects to the agents you already use; it is not affiliated with or endorsed by their makers. FROM YOUR PHONE • See every project and session on your computer, and run several sessions at once • Start a session, pick the agent and model, and give it its own Git worktree • Get a push notification when an agent finishes, asks a question or needs approval • Approve or reject permission requests where the agent supports them • Queue a follow-up while a turn is running, or stop it • Hold the mic and speak your prompt, then check the transcript before you send it • Attach images and use the agent's own slash commands and skills • Review the local Git diff, branch and pull request status PRIVATE BY DESIGN Your repository, toolchain and coding agent never move to a cloud workspace. The Sesori Bridge runs next to them on your computer and connects to your phone through a relay. Session traffic is end-to-end encrypted (X25519 key exchange, XChaCha20-Poly1305), so the relay passes ciphertext, not your code, prompts or AI responses. Voice transcription and notification previews use a separate, documented data path. HOW IT WORKS 1. Install Sesori Desktop on macOS or Linux and sign in. It runs the Bridge for you. On a server, over SSH or on Windows, run the headless Bridge CLI instead. 2. Sign in to this app with the same GitHub, Google, Apple or email account. 3. Your projects appear on your phone. Pick one and keep working. Not a mobile IDE and not a cloud sandbox: your agent runs where it always does, with your tools, files and settings. Your agent and model provider keep their own plans and pricing. Sesori is free. Setup guides and the security model: sesori.com and docs.sesori.com.

Your coding agents don't stop when you leave the desk, so neither should you. Tailscode puts every agent running on your own computers in your pocket: Claude Code, opencode and Oh My Pi, on any model they can run, including one on your own graphics card. Start a task from the couch, check on six at once from anywhere, and get a ping when the long one finishes. Your chats go straight to your machines over your own Tailscale network, never through a relay, with no sign-in. Free and open source. ONE COMMAND TO SET UP For Claude Code, one line on the computer you code on installs the bridge and ends with a QR code. Scan it, confirm, and the app has your machine; opencode has a one-line installer too. Want to look first? A built-in demo shows the main screens with no server at all. EVERY AGENT AT A GLANCE • Home is one ranked board: a sentence that says who needs you and what is working, then the chats waiting on you, working, pinned and recent. • The Agents board shows what needs you, what is working, what runs in the background and what finished in the last hour, with approve and answer on each card. Watch turns it into a full-screen wall of every agent. • Peek at any agent without opening its chat: its step, its last moves, its context. ANY MODEL, EVEN YOUR OWN GPU • Whatever your agent can run, from cloud models to a local model served by Ollama or llama.cpp on your own machine, picked per chat with its reasoning effort. • Carry a conversation to another model: the next one gets a short brief of the chat and picks up the thread. • A party of six favourite models and their effort levels, one tap or one swipe of the model pill away. LONG RUNS, NO BABYSITTING • A Live Activity on your Lock Screen for a running conversation, one for each with Pro: what it is doing, for how long, and how many tools it has run. • Notifications say what happened: how long it took, what it changed and the first of what it said. Reply from the Lock Screen to start the next turn. An approval shows the exact command and asks for Face ID or your passcode; a question shows its options as buttons. • Queue the next message while a turn runs, edit it before it goes, or stop the turn on the server. A NATIVE TRANSCRIPT, NOT A TERMINAL • Megatron folds a turn's thinking and tool runs into one head you open to read, search and replay every step. • Diffs, commands and their real output inline; syntax-highlighted code with byte-exact copy. • Send screenshots and photos; every picture the agent hands back lands in a gallery. NOT ONLY CODE Your machines know things. Ask what is eating the GPU, research a question, tidy a folder, check a price: any agent, any model, from anywhere. KNOW WHAT IT COSTS • Every conversation priced, every month totalled across machines, by model, project and tool. Always an estimate, never a bill. • Plan limits and quotas per provider and account, with the pace and turns left, on Home and in widgets. YOURS, END TO END • Your chats travel point to point over your tailnet, never through a relay. No sign-in, no ads, no tracking, and your code and conversations never reach us. • iPhone, iPad and Mac, plus a Linux desktop app, all open source under GPL-3.0: the app, the engine and the bridges. FREE, WITH AN HONEST PRO The free app is the complete core on one server: full chat, the Agents and Home boards, approvals, attachments, every model, a Live Activity at a time. Tailscode Pro is a one-time purchase, no subscription: unlimited servers in one list, concurrent Live Activities, iPhone and Mac on the same Apple Account, the delegate board, and a supporter heart. Tailscode is independent open-source software, not affiliated with or endorsed by Anthropic, Tailscale Inc., the opencode project or Oh My Pi. Privacy policy: https://midgarcorp.cc/tailscode/privacy Source code: https://github.com/guitaripod/Tailscode