PixCLI gives any AI agent the power to generate images, edit visuals, and produce videos. Headless. Self‑improving. One command.
npm install pixcli$ pixcli image "a neon cityscape at night, cyberpunk" --ratio 16:9
⠋ Generating...
✓ Created a-neon-cityscape-at-night-cyberpunk.png (1024×576, 2.1s)
$ pixcli edit "remove background, add transparency" -i ./cityscape.png
✓ Created remove-background-add-transparency.png (1024×576, 1.8s)
$ pixcli image "product shot of headphones" --quality high --count 2
✓ Created product-shot-of-headphones_1.png (1024×1024, 3.4s)
✓ Created product-shot-of-headphones_2.png (1024×1024, 3.4s)
$ pixcli video "30s product demo of the headphones" --fps 30
⠋ Rendering frames...
✓ Created product-demo-of-the-headphones.mp4 (1920×1080, 30fps, 12.7s)
4 tasks completed · 0 errors · 4 images · 1 video generated
Everything your agent needs to go creative. Nothing it doesn't.
No browser. No display server. No GPU. Runs anywhere your agent runs — CI, containers, serverless, your laptop.
Verifies its own output. Iterates automatically. Your agent learns from what it creates and gets better every time.
Generate images. Edit photos. Produce videos. All from a single CLI that your agent already knows how to use.
One install. Works immediately. No API keys to juggle, no dependencies to wrangle. It just works.
No more duct tape. No more setup guides. Plug it in. Ship creative work. Period.
Plug & play with
Claude Code
Anthropic
OpenAI Codex
OpenAI
OpenClaw
Open source
Hilo
Hosted OpenClaw
Paste this into Claude Code, Codex, OpenClaw, or any agent. It installs the CLI and teaches your agent how to use it.
Install PixCLI — the creative toolkit for AI agents. PixCLI lets you generate images, videos, voiceover, music, and sound effects from the command line. No GPU or browser needed. It works with any coding agent that supports SKILL.md (Claude Code, Codex, Gemini CLI, Cursor, and more). ## Step 1: Install the CLI npm install -g pixcli ## Step 2: Install the agent skill Download and unzip the skill pack into your agent's skills directory: curl -fsSL https://pixcli.hilo.cx/pixcli-skill.zip -o /tmp/pixcli-skill.zip && mkdir -p .agents/skills/pixcli && unzip -o /tmp/pixcli-skill.zip -d .agents/skills/pixcli && rm /tmp/pixcli-skill.zip This installs SKILL.md plus Remotion templates, 30+ reference docs, and a prompt cookbook. Alternative paths depending on your agent: - Claude Code: ~/.claude/skills/pixcli/ (personal) or .claude/skills/pixcli/ (project) - OpenClaw: clawhub install pixcli - Codex / Gemini CLI / Copilot: .agents/skills/pixcli/ (default above) - Cursor/Windsurf: copy SKILL.md content into .cursorrules or .windsurfrules ## Step 3: Authentication Set your Meterkey API key: export METERKEY_API_KEY="mk-..." Get your key at https://hilo.cx — one key covers all capabilities (images, video, voice, music, SFX). In Hilo/OpenClaw it's injected automatically as the METERKEY_API_KEY secret. ## Step 4: Verify pixcli --version pixcli image "a red circle on white background" -o test.png After setup, read the SKILL.md in the installed skill directory for the full command reference with all models, options, and creative workflows. Always use --json when calling pixcli programmatically.
npm install -g pixcli — one command, available everywhere.
Download the agent skill pack — Remotion templates, 30+ reference docs, and prompt cookbook.
Generate images, videos, voiceover, music, and sound effects — all from the terminal.
Be the first to give your agents creative superpowers.