Generate images, video & audio from your terminal.
Built to be run by your coding agent. Paste the prompt below into Claude Code, Cursor, or Codex — you're generating in seconds.
Set up the ploof CLI and generate assets for me: 1. Install: `bun i -g @miketromba/ploof` 2. Run `ploof learn` and follow it. 3. Make sure I'm logged in (`ploof whoami`); if not, help me. Then ask what I'd like to create — images, video, audio, logos, edits — suggest a few ideas, and make it.
Real assets. Real prompts.
Every image here was generated by Ploof — different styles, sizes, and subjects from one CLI. Hover any tile to see the exact prompt.
Brand mascot
--prompt “Adorable 3D character of a fluffy white cloud with a cheerful face and tiny arms holding a small glowing retro computer terminal, floating above soft pastel clouds, lavender and peach gradient background, gentle sparkles, octane render.”
Miniature city
--prompt “Wide cinematic tilt-shift photograph of a miniature paper city at blue hour, tiny warm glowing windows, soft fog drifting between paper skyscrapers, dreamy shallow depth of field.”
Product photo
--prompt “Studio product photograph of a matte black insulated water bottle on white marble, soft natural daylight, subtle reflections, premium minimal e-commerce look, shot on 50mm.”
Editorial portrait
--prompt “Editorial portrait of a woman with freckles, dramatic warm side lighting, soft film grain, muted earthy tones, shallow depth of field, shot on 85mm.”
App icon
--prompt “3D isometric app icon of a friendly robot mascot made of soft matte clay, pastel lilac and mint, rounded squircle, subtle studio shadow, clean light background.”
Vector illustration
--prompt “Flat vector illustration of a happy developer at a desk surrounded by floating UI panels and sparkles, soft pastel palette, clean rounded shapes, friendly modern tech brand style.”
Food flat lay
--prompt “Overhead flat lay photograph of a flat white coffee with latte art and a flaky croissant on a natural linen tablecloth, soft morning light, minimal styling, warm tones.”
3D abstract
--prompt “Abstract glossy 3D gradient blob, smooth liquid iridescent surface, lavender to peach, soft studio lighting on a clean off-white background, minimal, high detail.”
Need motion or sound? See video & audio ↓
Video and audio, too.
The same CLI generates motion and sound — the clip and voiceover below are straight out of Ploof.
“Slow cinematic dolly push through a miniature paper city at blue hour, tilt-shift, tiny glowing windows, soft drifting fog.”
ploof video generate --prompt "…" --seconds 4 --out city.mp4 “Ploof turns a prompt into a finished asset. Hand it to your agent, and watch your ideas show up as files.”
ploof audio generate --text "…" --voice alloy --out vo.mp3 One file. Many assets. One command.
Declare every asset in a YAML manifest and let Ploof run them in parallel. The gallery above was produced by exactly this.
- Mixed modalities — images, video, and audio in one run
- Dependency-aware: chain a task's output into the next with needs
- Runs concurrently; --dry-run validates before you spend a cent
ploof run gallery.yaml --parallel 4# gallery.yaml — build the whole gallery in parallel
version: 1
parallel: 4
tasks:
- id: mascot
kind: image.generate
prompt: "A fluffy cloud holding a glowing terminal…"
params: { model: gpt-image-2, size: 1024x1024, quality: high }
output: images/mascot.png
- id: city
kind: image.generate
prompt: "Tilt-shift miniature city at blue hour…"
params: { model: gpt-image-2, size: 1536x1024 }
output: images/city.png
- id: voiceover
kind: audio.generate
text: "Ploof turns a prompt into a finished asset…"
params: { model: gpt-4o-mini-tts, voice: alloy }
output: audio/voiceover.mp3
# …and 6 more imagesA small, sharp tool that does the boring parts.
No SDK wiring, no polling loops, no glue code. Just prompts in, files out.
Agent-native
Built to be driven by coding agents. ploof learn self-documents the installed version; output is clean JSON/JSONL with predictable exit codes.
Every modality
Images, video, and audio — generate, edit, extend, transcribe, and translate, all behind one consistent command surface.
Multi-provider
OpenAI today, plus fal.ai's entire model marketplace via model run. The provider registry is built to keep growing.
Batch & parallel
Declare assets in a YAML manifest, wire dependencies with needs, and run them concurrently with a single command.
Local auth profiles
Keep multiple keys per provider in ~/.ploof, switch with --profile, or override from environment variables in CI.
Reproducible
Every asset gets a <file>.json sidecar recording the prompt, params, and provider metadata — so any result is repeatable.
Two ways to start.
Let your agent do it (the usual path), or drive the CLI yourself.
With your coding agent
Paste the prompt, then tell it what to make. It installs Ploof, reads
ploof learn, and handles the rest.
Set up the ploof CLI and generate assets for me: 1. Install: `bun i -g @miketromba/ploof` 2. Run `ploof learn` and follow it. 3. Make sure I'm logged in (`ploof whoami`); if not, help me. Then ask what I'd like to create — images, video, audio, logos, edits — suggest a few ideas, and make it.
By hand
Prefer to drive it yourself? Three commands to your first asset.
# 1 — install
bun i -g @miketromba/ploof
# 2 — authenticate (stored in ~/.ploof)
ploof login openai --api-key sk-...
# 3 — generate your first asset
ploof image generate \
--prompt "Studio product photo of a matte black bottle" \
--out hero.png
Ready to make something?
Drop the prompt into your agent and describe what you want. Files in seconds.