kitty 0.50 ships docked windows, session restore

// 1h agoPRODUCT UPDATE

kitty 0.50 ships docked windows, session restore

kitty 0.50 adds docked windows for status bars, file trees, and adjacent shells, plus opt-in restoration of windows, tabs, and scrollback after logout or reboot. It also introduces workspace-like tab targeting and improves Wayland desktop integration. [Changelog](https://sw.kovidgoyal.net/kitty/changelog/)

// ANALYSIS

This is a workflow-focused release that makes kitty more compelling as a programmable desktop workspace, especially for developers managing persistent local or AI coding sessions.

–Docked windows add practical layouts for status panels, file trees, and always-visible shells without consuming normal split space.
–Session restoration preserves terminal context, including scrollback; macOS encrypts saved state, while Linux requires the Inhibit desktop portal.
–Named-tab controls and `launch --focus-existing` improve automation across multi-project workflows.
–Wayland decoration improvements and touchscreen support make kitty feel more native on modern Linux desktops.
–The update is not AI-specific, but its scripting, session, and composability features suit terminal-based coding agents. [Product homepage](https://sw.kovidgoyal.net/kitty/)

// TAGS

kittyclidevtoolopen-sourcelocal-first

DISCOVERED

1h ago

2026-11-04

PUBLISHED

5h ago

2026-11-04

RELEVANCE

5/ 10

AUTHOR

porridgeraisin

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE52m ago

MailCheap Builds Momentum With Early Testimonials

MailCheap’s week-three update highlights its first customer testimonials, validating a newsletter platform positioned as up to 83% cheaper than beehiiv. The product charges by emails sent and combines open-source self-hosting with API and MCP access for AI agents.

VIDEO1h ago

Fable 5.5 Renders 66-Second Film in Code

An X demo claims Anthropic’s unreleased Fable 5.5 autonomously produced a 66-second animation—including visuals, music, and sound effects—using Python-generated code alone. The result is compelling, but Anthropic has not confirmed the model and its provenance remains unverified.

RESEARCH1h ago

NVIDIA Paper Questions Need for Long Context

NVIDIA’s technical report finds that frequent context compaction preserves accuracy across many long-horizon and long-input tasks while reducing active context costs. It also introduces semantic state space models, where agents carry forward compact, structured state instead of full conversation history.