Feature request: smooth/paced text reveal in fullscreen TUI instead of instant chunk rendering

Feature description

When Claude streams a response, text currently appears in instant bursts as chunks arrive from the API. Each repaint pops a block of text into place with no transition, which feels jarring — especially in long responses where several paragraphs can land at once.

Request: an opt-in setting to meter streamed text out at a smooth, steady rate (a paced "typewriter-style" reveal), similar to how the claude.ai web app animates responses in. The fullscreen TUI renderer (tui: "fullscreen" / CLAUDE_CODE_NO_FLICKER=1) already owns every repaint and virtualizes scrollback, so it seems like the natural place to buffer incoming chunks and drain them at a fixed characters-per-frame rate instead of printing them the moment they arrive.

Suggested shape

• A setting like "textRevealRate" (chars/sec, or "instant" to keep current behavior — default)
• Reveal rate should adapt so the animation never falls meaningfully behind the actual stream (catch up if the buffer grows large)
• Instant flush on turn completion, Ctrl+C, or when the user scrolls up

Context

• Fullscreen mode already fixed the flicker side of this; this is the remaining "content pops in" half.
• Terminal emulators can't solve it themselves: e.g. Ghostty custom shaders only receive the current frame plus cursor uniforms (no feedback buffer), so fade-in can't be done at the terminal layer — it has to come from the app doing the writing.

Environment: Claude Code with tui: "fullscreen", Ghostty on macOS (arm64).

🤖 Filed with Claude Code on behalf of the user.