Persist chat tools and stream artifact drafts

Make reasoning preview behave like a terminal line

Collapse live reasoning and suppress warm prepare

Clarify warm inference preparation

Flatten artifact file tool arguments

Keep model resident on cancel and harden artifact tools

Persist tool logs and recover chats with invalid artifacts

Fix artifact tool termination and dynamic transcript scrolling

Remove duplicate shard verification pass

Warm the active chat prompt shape

Warm WebGPU before first prompt

Use manifest context limits and handle memory exhaustion safely

Set 8K context defaults and add artifact ZIP export

Add persistent multi-file artifact workspaces

Render ordered agent flow and edit messages

Accept expression input in js_eval

Add agentic artifact tools and live diagnostics

Add resizable sidebars and inline artifacts

Ship validated WebGPU Q1 turbo runtime

Assert 4096-token runtime defaults

Use 4096-token release defaults

Fix streamed tool calls and align generation limits

Release v1.2.2 interactive chat workspace

Release v1.2.1 adaptive chat workspace

Release v1.2.0 live model telemetry

Release v1.1.4 manifest progress fix

Release v1.1.3 scored WebGPU evidence

Release v1.1.2 benchmark runtime evidence

Select the flagship 27B tier by default

Add static deep-link benchmark route and release audit fixes

Add Phase 5 benchmark and runtime provenance

Release browser-local Bonsai WebGPU chat