Paste one prompt into Claude Code, Codex, Cursor, Gemini CLI or any agent that can run a command. It installs the FaceNinja skill, signs you in with a code you approve in the browser, and makes the head. No key to copy.
Sign up with your email or Google. The first head is on us.
Your agent reads the skill, installs it and starts the sign-in.
The agent shows a code like WXYZ-2345. Open the link, check the code, click Approve. The agent gets its own API key and makes the head.
Put a photo next to where your agent works and name it photo.jpg, or change the path in the prompt. None of these prompts contains a key.
Read https://faceninja.net/skill/SKILL.md and install the FaceNinja skill it describes. Then sign me in with it (it will show me a code to approve) and make a MetaHuman head from ./photo.jpg. Save the GLB and the MetaHuman ZIP in ./faceninja-out.
Works in any agent that can fetch a URL and run a shell command.
Install the FaceNinja skill with `curl -fsSL https://faceninja.net/skill/install.sh | sh -s -- --claude`, then read the SKILL.md it installed and follow it: sign me in (show me the code to approve) and make a MetaHuman head from ./photo.jpg. Save the GLB and the MetaHuman ZIP in ./faceninja-out.
Installs to ~/.claude/skills/faceninja (or $CLAUDE_CONFIG_DIR/skills). New sessions load it by itself.
Install the FaceNinja skill with `curl -fsSL https://faceninja.net/skill/install.sh | sh -s -- --codex`, then read the SKILL.md it installed and follow it: sign me in (show me the code to approve) and make a MetaHuman head from ./photo.jpg. Save the GLB and the MetaHuman ZIP in ./faceninja-out.
Installs to ~/.codex/skills/faceninja (or $CODEX_HOME/skills). The sign-in and downloads need network access.
Install the FaceNinja skill into this project with `curl -fsSL https://faceninja.net/skill/install.sh | sh -s -- --cursor`, then read .cursor/skills/faceninja/SKILL.md and follow it: sign me in (show me the code to approve) and make a MetaHuman head from ./photo.jpg. Save the GLB and the MetaHuman ZIP in ./faceninja-out.
Installs to .cursor/skills/faceninja in the open project.
Install the FaceNinja skill into this project with `curl -fsSL https://faceninja.net/skill/install.sh | sh -s -- --agents`, then read .agents/skills/faceninja/SKILL.md and follow it: sign me in (show me the code to approve) and make a MetaHuman head from ./photo.jpg. Save the GLB and the MetaHuman ZIP in ./faceninja-out.
Installs to .agents/skills/faceninja. Add --agents-md to also list it in AGENTS.md. Other agents use the same line.
# A key: Factory → API keys, or `python3 faceninja.py login` export FACENINJA_API_KEY=fn_live_... # 1. create a head from a photo (held: 1,500 credits) curl -s https://faceninja.net/v1/heads -H "Authorization: Bearer $FACENINJA_API_KEY" \ -H "Idempotency-Key: my-first-head" -F photos=@photo.jpg -F name=Me # 2. wait (repeat until "status" is "completed" or "failed") curl -s "https://faceninja.net/v1/heads/hd_...?wait=60" -H "Authorization: Bearer $FACENINJA_API_KEY" # 3. download (the link works for 60 seconds and needs no key) curl -s https://faceninja.net/v1/heads/hd_.../files -H "Authorization: Bearer $FACENINJA_API_KEY" -d format=glb curl -s https://faceninja.net/v1/heads/hd_.../files -H "Authorization: Bearer $FACENINJA_API_KEY" -d format=metahuman_zip
Full reference: https://faceninja.net/docs/
Use my FaceNinja API key fn_live_YOUR_KEY as the environment variable FACENINJA_API_KEY for the commands you run; do not write it into any file. Read https://faceninja.net/skill/SKILL.md, install the FaceNinja skill it describes and make a MetaHuman head from ./photo.jpg. Save the GLB and the MetaHuman ZIP in ./faceninja-out.
This prompt contains a live API key. Agent transcripts and logs keep it: anyone who reads them can spend your credits. Prefer the sign-in prompt, and revoke this key in Factory → API keys when you are done.
The skill is a folder in the open Agent Skills format: SKILL.md, one Python script and short API notes. The installers download it, check each file's SHA-256, copy it into the skills folder of every agent they find and say where. They change nothing else.
curl -fsSL https://faceninja.net/skill/install.sh | sh
Add -s -- --claude, --codex, --cursor or --agents to choose the folder. Read the script first if you like.
irm https://faceninja.net/skill/install.ps1 | iex
Read the script. The skill needs Python 3.9 or newer.
| Claude Code | ~/.claude/skills/faceninja (or $CLAUDE_CONFIG_DIR/skills) |
|---|---|
| Codex | ~/.codex/skills/faceninja (and $CODEX_HOME/skills) |
| Cursor | .cursor/skills/faceninja in the open project |
| Gemini CLI and others | .agents/skills/faceninja, plus one line in AGENTS.md if you agree |
Files: SKILL.md · faceninja-skill.zip · faceninja.py (the script alone) · api.md · version.json (checksums) · llms.txt
The skill drives faceninja.py, a single file with no dependencies. People can run it too.
login: device sign-in, the key saved readable by you onlybalance, whoami: credits left, who is signed inmake: 1 to 15 photos of one person to a head, then GLB, FBX, OBJ, STL, PLY or the MetaHuman ZIPbatch: a folder, one head per photo or per subfolder, resumablestatus, download, list, logout--json on every command, exit codes an agent can act onBuilding your own integration? The API reference covers the same workflow with curl, Node and Python, and the AI agents section lists the sign-in calls.