Nothing here yet — your generations will show up live.
Gallery
0 selected
Agent skill
teach another AI agent to generate images with this server
qwen-image-studio — text to image & image edit
One Markdown file that teaches any AI agent (pi, Claude Code, Codex, Cursor, or a plain
chat model) to drive the t2i and edit endpoints of this studio over plain
HTTPS — hand it the key, submit, poll, download. No sign-up, no SDK, no dependencies.
A key is a bearer token: send Authorization: Bearer h3sk_… on any request and
the agent needs no password. It can generate media and manage that account’s files,
but never accounts or other keys. Click a key to reveal it.
1
Unzip into a skills folder (pi, Claude Code)
2
Or paste it into a system prompt / AGENTS.md (any other agent)
Copy SKILL.md above and add one line with the address and the account to log in as.
3
Then ask it for a picture
“Use the qwen-image-studio skill to make a 3:2 photo of a lighthouse in fog”, or
“edit photos/x.jpg: make it night, keep the face”. It runs the CLI, waits, and saves the PNG.