Whiskers gives you full control of your AI with memory, continuity and local‑first architecture.
Whiskers summarises older messages to maintain a long memory without bloating the context window. You can manually summarise or rewrite responses to shape the conversation using commands like /rewrite, /context or /summarize.
Guide the AI’s behaviour with built‑in commands and custom contract files. Adjust canon capsules on the fly to define characters, worlds or tasks, and rewrite responses in real time to get the story you want.
Sessions persist independently of browser windows. Start a multi‑day campaign and pick up exactly where you left off without losing context.
Your conversations never leave your computer. Whiskers runs in your browser while using a local llama.cpp model server. Nothing is sent to the cloud.
Choose and run your own model on your machine. Whiskers talks to a local server so you retain full ownership and there’s zero network latency.
