55
66** A live visual surface for your terminal coding agent.**
77
8- Your agent works in a wall of text. sideshow gives it a screen. It publishes
8+ Your agent works in a wall of text; Sideshow gives it a screen. It publishes
99** surfaces** — diagrams, UI sketches, rendered markdown, syntax-highlighted
1010diffs, terminal output, images — and they render live in your browser while it
11- works. You comment back, and your comments reach the agent. Publish, render,
12- comment, revise: a two-way loop, right next to the conversation.
11+ works.
1312
1413<table >
1514 <tr >
@@ -30,14 +29,13 @@ comment, revise: a two-way loop, right next to the conversation.
3029- ** See what your agent means.** An architecture it's proposing, a flow it's
3130 tracing, a UI it's about to build — shown, not described in a paragraph you
3231 have to picture in your head.
32+ - ** Multimodal.** Combine diffs with mermaid diagrams, terminal output with HTML
33+ explainers, and more. Combine surfaces to illustrate ideas better.
3334- ** Faster, and fewer tokens.** sideshow surfaces are optimized for LLMs, so the
3435 agent produces higher-fidelity diagrams and visualizations faster and using
3536 fewer tokens.
36- - ** Comment instead of re-prompting.** Click a card, leave a note. The agent
37- reads it and revises — no copy-pasting screenshots or re-describing what you
38- meant.
39- - ** Works with the agent you already use.** Any agent with a shell can drive it
40- over ` curl ` ; richer tiers exist for CLI, MCP, Pi, and Claude Code.
37+ - ** Works with the agent you already use.** Works with any agent harness: Claude Code
38+ (desktop or CLI), Codex (desktop or CLI), Opencode, Pi, etc.
4139
4240## Quick start
4341
0 commit comments