Every release, in plain language.

Wayland 0.11.8 leads with a browser and computer-use overhaul: built-in browser capability that opens pages, clicks, and fills forms, plus a one-pass macOS permission flow for sharper computer use. Add response-length control you set, a persistent per-project workspace, and clean long sessions across every provider, all on the Wayland Core v0.12.17 engine.

Wayland 0.11.7 introduces Concierge, the first built-in, setup-aware guide. It reads your own Wayland and answers in plain English, then sets things up for you: it proposes a change, you click Apply, done. Plus suggested starting cards on every assistant, a steadier Flux Auto reasoning budget, draft recovery, and smarter macOS auto-update, all on the new Wayland Core v0.12.16 engine.

Wayland 0.11.5 makes local models keyless. Point at a local Ollama or llama.cpp endpoint and run fully private and free, with zero keys, on a cleaner, tidier sidebar, all on the new Wayland Core v0.12.15 engine.

Wayland 0.11.4 ships Crucible, a Mixture-of-Providers council that convenes multiple models on one task and fuses the strongest answer, with Flux Auto routing, greatly improved chat streaming, real sandboxed tool-use on Windows, and local models that pull real weight.

Wayland 0.11.3 sends every connected key to the right provider at spawn, welcomes Sakana AI as a first-class provider, streams live model catalogs, runs local Ollama keyless, and wires global Wayland Memory into chat.

A redesigned provider picker, native routing across 104 providers, friendly model names everywhere, an in-app engine updater, Flux Voice, and one-click migration from other tools.

An MCP marketplace of 107 integrations, image generation in chat, a runaway-loop safety net, one-click Flux failover, phone-based setup, and Operator Teams.

Sign in with X, browser-style tabs and pop-out chats, a safer Autopilot, built-in Doctor, custom slash commands, and a full mobile pass.