Version 2026.4.23 (Before they started screwing everything up)

The read tool has been modified in this release to add images, audio, or video to chat. Simply add to TOOLS.md and tell AI model to read it to understand the command. See example TOOLS-EXAMPLE.md

This custom build adds comprehensive media support to OpenClaw across multiple channels:

🌐 WebChat (Complete media suite)

Previously had NO media support - now fully featured:

Images:

• JPEG/JPG
• PNG
• GIF
• WebP
• SVG
• BMP

Audio (local files - streamed via HTTP server on port 18791):

• OGG (.ogg) - audio/ogg
• MP3 (.mp3) - audio/mpeg
• WAV (.wav) - audio/wav
• FLAC (.flac) - audio/flac
• M4A (.m4a) - audio/mp4
• AAC (.aac) - audio/aac
• Opus (.opus) - audio/opus
• WebM audio (.webm) - audio/webm
• WMA (.wma) - audio/x-ms-wma

Video (local files - streamed via HTTP server on port 18791):

• MP4 (.mp4) - video/mp4
• WebM (.webm) - video/webm
• AVI (.avi) - video/x-msvideo
• MOV (.mov) - video/quicktime
• MKV (.mkv) - video/x-matroska
• M4V (.m4v) - video/x-m4v
• MPG/MPEG (.mpg, .mpeg) - video/mpeg

Social Media Embeds: (Disabled by default)

• YouTube
• Twitter/X
• TikTok
• Instagram
• (Other platforms via URL detection)

📱 Telegram

Images:

• JPEG/JPG
• PNG
• GIF
• WebP
• SVG
• BMP

Platform summary

Platform	Images	Audio	Video	Social Embeds
WebChat	✅	✅	✅	✅
Telegram	✅	❌	❌	❌

Extra features

• Added resize handles on webchat messages so you can resize media!
• Added audio/video memory playback. Close or refresh the browser, it remembers where you stopped.
• Added new core tool for downloading YouTube and other video's called download_video using yt-dlp
• UI changes to get rid of most of the dead screen space. Larger area now for messages.

Implementation details

• Local HTTP media server (port 18791) for WebChat streaming of local files.
• Extension-based MIME mapping for audio/video
• MIME detection for images
• Social media URL pattern matching

Breaking changes

None - all features are additive.

Some Screenshots

https://github.com/user-attachments/assets/a11f41b2-0262-48c7-8ba2-7c5462840d6f

Examples

See TOOLS-EXAMPLE.md
See openclaw-EXAMPLE.json

🦞 OpenClaw — Personal AI Assistant

Install

This install/configuration has been tested running under Ubuntu 24 under WSL under Windows 10

I recommend you install the app globally and then have your config in a user folder like /home/openclaw (See releases for user config portion)

To build and install the app globally, simply run from root folder: ./global
sudo is required to install globally so you will be asked for your sudo password if you install globally.
This command will build the release and then install for all users under Ubuntu.

OR (Build manually, run setup manually, and build and run under that folder. Commands are different to run the build like this, just Google search or look at main OpenClaw release)

git clone https://github.com/jdc4429/openclaw.git
cd openclaw

pnpm install

# First run only (or after resetting local OpenClaw config/workspace)
pnpm openclaw setup

pnpm ui:build

# Dev loop (auto-reload on source/config changes)
pnpm gateway:watch

pnpm build

pnpm openclaw setup writes the local config/workspace needed for pnpm gateway:watch. It is safe to re-run, but you normally only need it on first setup or after resetting local state. pnpm gateway:watch does not rebuild dist/control-ui, so rerun pnpm ui:build after ui/ changes or use pnpm ui:dev when iterating on the Control UI. If you want this checkout to run onboarding directly, use pnpm openclaw onboard --install-daemon.

Note: You still need to install the configuration portion in a user folder ie. /home/openclaw unless you run from under the build

I created a copy with an example config you can simple download and 'tar xzvf file.tgz ~/' to extract to your current user folder.
The user config extracts to the .openclaw directory. (Hidden unless you use ls -la) Edit openclaw.json config to add your API keys, change root folders, and run setup for some extra plugins.