Video & Audio Generation

Wunderland includes built-in tools for generating video, animating images, analyzing video content, and creating music and sound effects. All features are accessible from the CLI and available as tools in wunderland chat.

---

Video Generation

Generate a video from text

wunderland video generate "aerial flyover of a neon-lit cyberpunk city at night"

The CLI sends the prompt to the configured video provider (Runway Gen-3 or Fal.ai), polls for completion, and saves the resulting MP4 locally.

Animate a still image

wunderland video animate ./photo.png "slow zoom out, camera pans right, cinematic lighting"

Image-to-video takes a source image and a motion description. Useful for creating short clips from generated artwork or product images.

Analyze a video

wunderland video analyze ./clip.mp4

Runs scene detection, object recognition, action classification, and sentiment analysis using a vision-capable LLM. Returns structured JSON with timestamps, descriptions, and confidence scores.

# Override the vision model
wunderland video analyze ./clip.mp4 --model gpt-4o

Flags

Flag	Effect	Default
--provider <name>	Video provider: runway, fal	First available
--model <name>	Provider-specific model override	Provider default
--duration <seconds>	Target video duration	5
--output <path>	Output file path	Auto-generated in current directory

---

Music Generation

wunderland audio music "lo-fi hip hop beat, rainy day vibes, 90bpm"

Generates a music track from a text prompt using Suno or Fal.ai.

# Specify provider and duration
wunderland audio music "epic orchestral battle theme" --provider suno --duration 60

# Custom output path
wunderland audio music "calm piano ambient" --output ./background.mp3

---

Sound Effect Generation

wunderland audio sfx "door creaking open slowly in an empty hallway"

Generates a short sound effect using Stable Audio or Fal.ai. Default duration is 5 seconds.

# Longer sound effect
wunderland audio sfx "thunderstorm building over 15 seconds" --duration 15

# Use a specific provider
wunderland audio sfx "mechanical keyboard typing" --provider stability

---

Provider Configuration

Set default providers so you do not need to pass --provider every time:

wunderland extensions configure

Or set them directly:

wunderland config set videoProvider runway
wunderland config set musicProvider suno
wunderland config set sfxProvider stability

Provider Matrix

Capability	Providers	Env Var
Video generation	Runway Gen-3, Fal.ai	RUNWAY_API_KEY, FAL_API_KEY
Image animation	Runway Gen-3, Fal.ai	RUNWAY_API_KEY, FAL_API_KEY
Video analysis	OpenAI (GPT-4o), Anthropic (Claude Sonnet), Google (Gemini Pro)	LLM key with vision support
Music generation	Suno, Fal.ai	SUNO_API_KEY, FAL_API_KEY
Sound effects	Stable Audio, Fal.ai	STABILITY_API_KEY, FAL_API_KEY

---

Using in Chat

All video and audio tools are available as agent tools in interactive chat sessions:

wunderland chat
> Generate a 10-second video of waves crashing on a rocky shore at sunset
> Create a lo-fi beat to go with it
> Analyze this video file: ./demo.mp4

The agent calls generate_video, generate_music, analyze_video, etc. as tool invocations. When selfImprovement is enabled the agent can learn your preferred providers and styles over time.

---

Combining with Provider Preferences

Use providerPreferences in agent.config.json to control which provider handles each media type. See Provider Preferences for details.

{
  "providerPreferences": {
    "video": { "preferred": ["runway"], "blocked": [] },
    "music": { "preferred": ["suno"], "blocked": [] },
    "sfx": { "preferred": ["stability"], "blocked": ["fal"] }
  }
}