PR Submitted: Inline Image Rendering in the OpenClaw TUI (iTerm2, WezTerm, Ghostty)

A community member just submitted a pull request that brings inline image rendering to the OpenClaw terminal UI. When a tool call returns an image, instead of just printing the file path, the image now displays directly in your terminal.

How It Works

The PR leverages the iTerm2 Inline Images Protocol — a widely-supported standard that allows terminals to render actual pixel graphics inline with text output. When your agent returns an image from a tool call (screenshots, generated charts, camera snapshots, etc.), the TUI now detects this and outputs the image data using the appropriate escape sequences.

Supported Terminals

The feature works in terminals that implement the iTerm2 inline image protocol:

• iTerm2 — The original (macOS)
• WezTerm — Cross-platform, Rust-based
• Ghostty — The new hotness from Mitchell Hashimoto
• Kitty — Has its own protocol but often supports iTerm2 sequences too
• Konsole — KDE's terminal (partial support)

If your terminal doesn't support the protocol, the behavior falls back to printing the file path as before — no breakage.

Why This Matters

For those of us who live in the terminal, this is a significant quality-of-life improvement. Consider the workflow:

1. You ask your agent to take a screenshot of a webpage
2. The browser tool captures the image
3. Before: You see /tmp/screenshot-abc123.png and have to open it separately
4. After: The screenshot appears right in your terminal, inline with the conversation

This is especially useful for:

• Visual debugging — See what the browser automation captured without leaving your terminal
• Camera snapshots — Quick visual checks from node-hosted cameras
• Generated content — Charts, diagrams, or any visual output from tools
• Faster iteration — No context-switching to a separate image viewer

The Technical Bits

The iTerm2 protocol uses OSC (Operating System Command) escape sequences to embed base64-encoded image data. The terminal interprets these sequences and renders the image inline. The PR handles the encoding and formatting, so the TUI just works.

Try It Out

If you're comfortable building from source, you can test the PR branch. Otherwise, keep an eye on the next release — this seems likely to land given how clean the implementation is.

For terminal enthusiasts who prefer the CLI over the web UI, this brings OpenClaw's visual capabilities into your native environment. No more hunting for file paths.