Skip to content

v0.11.4

Pre-release
Pre-release

Choose a tag to compare

View all tags
@github-actions github-actions released this 28 Jun 09:29
v0.11.4
42b3034
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Stop picking the best AI. Command all of them.

The real unlock was never choosing the best model. It's never having to choose one at all.

Wayland 0.11.4 runs every major AI side by side and fields the right one for the job automatically. And it now ships on a new engine, Wayland Core 0.12.14 with Crucible, our Mixture-of-Providers council, so the hardest problems can be worked by a whole panel of providers at once instead of a single model.

✨ Highlights

  • Mixture-of-Providers, powered by Crucible. The bundled engine now packs Crucible, a cross-provider council that can convene multiple models on one task, let each take a swing, and fuse the strongest answer. In-app, Flux Auto routes each request to the best model automatically, so you stop guessing which AI to use.
  • Greatly improved chat streaming. Hit send and watch it work: an inline orbit indicator that animates while the assistant runs and settles when it's done, a unified plain-English activity timeline across every backend, live reasoning summaries in the thinking block, smooth fast streaming, a per-message action toolbar the moment a turn completes, and a turn timer that stays accurate even when you switch chats.
  • Real tool-use on Windows, safely sandboxed. Wayland runs real shell commands and tools on Windows inside a locked-down AppContainer sandbox, working on a clean non-developer machine, not just a developer's rig. Live-verified.
  • Your local models pull real weight. Ollama and llama.cpp models run private and free, and now handle tools that used to break them, thanks to up-front capability detection with a graceful no-tools fallback.

🔧 Also in this release

  • New engine Wayland Core 0.12.14: Crucible Mixture-of-Providers council, sharper cross-vendor cost and pricing accuracy, and the Windows AppContainer shell-tool fixes (grant-skip for missing dev caches, job-tree reap so commands return promptly).
  • Accurate, live model catalogs for every connected provider; assistants run on Wayland Core by default; Grok "Sign in with X" stays authenticated via a single, race-free OAuth refresher.
  • One-click assistants with starter-prompt grids, multi-model teams, and scheduled tasks.
  • Cleaner billing-error classification, a hardened model picker that never strands you on an empty list, and a stack of reliability fixes across the app. 63 changes in this release.

Free to start. One click. No setup ceremony.

Full technical notes: see CHANGELOG.md.

Assets 20
Loading