### Install genmedia CLI

Source: https://github.com/fal-ai-community/skills/blob/main/README.md

Installs the default skill bundle for projects with genmedia installed. This command sets up skills in `.agents/skills/` or `.claude/skills/`.

```bash
genmedia init
```

--------------------------------

### Run genmedia setup wizard

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Execute this command to launch the interactive setup wizard for genmedia configuration, including API key, output mode, and auto-update settings.

```bash
genmedia setup
```

--------------------------------

### Quick Code Example for Running GPT Image 2

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-prompting/references/gpt-image-2.md

Demonstrates how to execute the GPT Image 2 skill using the genmedia CLI. This example shows common parameters for prompt, image size, quality, number of images, output format, and download path.

```bash
genmedia run openai/gpt-image-2 \
 --prompt "<five-section structured prompt>" \
 --image_size landscape_4_3 \
 --quality high \
 --num_images 1 \
 --output_format png \
 --download "./out/{request_id}_{index}.{ext}" \
 --json
```

--------------------------------

### Install Fal Redesign Skill

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-redesign/README.md

Instructions for installing the fal-redesign skill and setting up the necessary environment variables.

```bash
ln -s /path/to/skills/fal-redesign ~/.claude/skills/fal-redesign
export FAL_KEY=... # https://fal.ai/dashboard/keys
```

--------------------------------

### Initialize Default Skill Bundle

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Installs the default genmedia skill bundle. Use --force to overwrite existing files.

```bash
genmedia init
```

```bash
genmedia init --force # overwrite existing files
```

--------------------------------

### GPT Image Generation v2 - Quick Code Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-prompting/references/gpt-image-2.md

A quick bash example demonstrating how to run the `openai/gpt-image-2` model with common parameters.

```APIDOC
## Quick Code Example

```bash
genmedia run openai/gpt-image-2 \
 --prompt "<five-section structured prompt>" \
 --image_size landscape_4_3 \
 --quality high \
 --num_images 1 \
 --output_format png \
 --download "./out/{request_id}_{index}.{ext}" \
 --json
```

For edits, swap to `openai/gpt-image-2/edit` and pass `image_urls` (array) plus optional `mask_image_url`.
```

--------------------------------

### Install genmedia on Linux/macOS

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Use this command to install the genmedia CLI on Linux and macOS systems by piping the installation script to bash.

```bash
curl https://genmedia.sh/install -fsS | bash
```

--------------------------------

### Install and Run fal-site

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-redesign/runtime/README.md

Set your Fal API key and run fal-site with a text prompt to generate website variations.

```bash
export FAL_KEY=... # https://fal.ai/dashboard/keys
npx fal-site "a calm Parisian coffee subscription, warm editorial feel"
```

--------------------------------

### Install genmedia on Windows

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Use this command to install the genmedia CLI on Windows systems using PowerShell.

```powershell
irm https://genmedia.sh/install.ps1 | iex
```

--------------------------------

### Install Genmedia

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/SKILL.md

Installs the genmedia tool. Use the Linux/macOS command for those systems and the Windows PowerShell command for Windows.

```bash
curl https://genmedia.sh/install -fsS | bash
```

```powershell
irm https://genmedia.sh/install.ps1 | iex
```

--------------------------------

### Image Processing Workflow Examples

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/references/utility-endpoints.md

These examples illustrate common image processing workflows using utility endpoints. Verify endpoint availability and inspect schemas before use.

```text
source image -> sam-hq mask -> invert or refine mask -> new background -> composite-image
```

```text
master asset -> resize 1:1 -> resize 9:16 -> compress-image
```

```text
video -> auto-subtitle -> optional setpts-video -> final download
```

```text
product image -> segment -> composite clean background -> image-grid -> compress
```

--------------------------------

### fal-site Usage Examples

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-redesign/runtime/README.md

Demonstrates different ways to use fal-site, including direct prompts, file input, and standard input.

```bash
npx fal-site "<free-form context>"
```

```bash
npx fal-site --file brief.txt -n 10
```

```bash
cat email.txt | npx fal-site --stdin -n 8 -o ~/Desktop/out
```

--------------------------------

### Discover Speech-to-Text Documentation

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-models-catalog/references/audio-to-text.md

Search for documentation related to speech-to-text. This command helps in finding relevant guides and examples for STT tasks.

```bash
genmedia docs "speech to text" --json
```

--------------------------------

### Setup Fal CLI and Project Directory

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-redesign/DEMO.md

Set your Fal API key and navigate to your web project directory. This is a prerequisite before running any redesign commands.

```bash
export FAL_KEY=... # https://fal.ai/dashboard/keys

cd ~/mon-projet-web # contains index.html (PixelFind)
open index.html # show the "before" in the browser

# Open a second terminal for the skill commands
cd ~/mon-projet-web
```

--------------------------------

### Setup Genmedia with API Key

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/SKILL.md

Sets up genmedia non-interactively using your FAL API key. Ensure the FAL_KEY environment variable is set.

```bash
genmedia setup --non-interactive --api-key "$FAL_KEY"
```

--------------------------------

### Get Full CLI Help in JSON

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Retrieves a machine-readable description of all commands, arguments, and options in JSON format.

```bash
genmedia --help --json
```

--------------------------------

### Install Runtime Dependencies

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-redesign/SKILL.md

Install the Node.js dependencies for the FAL skill runtime. Requires Node.js 18+.

```bash
cd runtime && npm install
```

--------------------------------

### Image-to-Video Prompt Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-prompting/references/happy-horse.md

When generating video from a still image, the prompt should only describe the desired motion and atmosphere. The reference frame provides identity, wardrobe, and lighting.

```text
slow camera push-in toward the subject, subtle wind, ambient haze
```

--------------------------------

### Minimal Workflow JSON Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/SKILL.md

This is a basic example of a fal.ai workflow JSON file. It defines a simple text-to-image pipeline with an input prompt and an output image. Ensure the 'app' field points to a valid fal.ai application.

```json
{
 "name": "my-workflow",
 "title": "My Workflow",
 "contents": {
  "name": "workflow",
  "nodes": {
   "output": {
    "type": "display",
    "id": "output",
    "depends": ["node-image"],
    "input": {},
    "fields": { "image": "$node-image.images.0.url" }
   },
   "node-image": {
    "type": "run",
    "id": "node-image",
    "depends": ["input"],
    "app": "fal-ai/flux/dev",
    "input": { "prompt": "$input.prompt" }
   }
  },
  "output": { "image": "$node-image.images.0.url" },
  "schema": {
   "input": {
    "prompt": {
     "name": "prompt",
     "label": "Prompt",
     "type": "string",
     "required": true,
     "modelId": "node-image"
    }
   },
   "output": {
    "image": { "name": "image", "label": "Generated Image", "type": "string" }
   }
  },
  "version": "1",
  "metadata": {
   "input": { "position": { "x": 0, "y": 0 } },
   "description": "Simple text to image workflow"
  }
 },
 "is_public": true,
 "user_id": "",
 "user_nickname": "",
 "created_at": ""
}
```

--------------------------------

### FAL Workflow Reference Syntax Examples

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/SKILL.md

Lists common reference syntaxes used in FAL workflows for accessing input fields and node outputs.

```markdown
| Reference | Use Case |
|-----------|----------|
| `$input.field` | Input value |
| `$node.output` | LLM text output |
| `$node.images.0.url` | First image URL |
| `$node.image.url` | Single image URL |
| `$node.video.url` | Video URL |
| `$node.audio_file.url` | Audio URL |
| `$node.frame.url` | Extracted frame |
```

--------------------------------

### Image-to-Video Motion Prompt Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-prompting/references/kling.md

When generating video from an image with Kling, the prompt should focus solely on motion. Avoid re-describing the static scene, as the reference frame already defines identity and appearance.

```text
slow camera push-in toward the subject, subtle wind in hair, ambient haze drift
```

--------------------------------

### Location Anchor Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/storytelling.md

Ensures scene consistency by defining room layout, window direction, color palette, and time of day.

```text
same room layout, window direction, color palette, and time of day
```

--------------------------------

### Campaign Manifest - Meta 4:5 Proof Still

Source: https://github.com/fal-ai-community/skills/blob/main/skills/marketing/references/examples.md

Use for Meta 4:5 proof stills. This example specifies the `openai/gpt-image-2` endpoint and requires a safe zone for external copy.

```text
Meta 4:5
`openai/gpt-image-2`
safe zone for external copy
```

--------------------------------

### Discover Video Editing Documentation

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-models-catalog/references/video-to-video.md

Use the genmedia CLI to find documentation related to 'video editing'. This command helps in locating relevant guides and information.

```bash
genmedia docs "video editing" --json
```

--------------------------------

### Kling O3 Text-to-Video Prompt Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-prompting/references/kling.md

Use this structure for single-shot text-to-video generation with Kling O3. Focus on direct, declarative descriptions of the subject, action, setting, time, camera, and lighting.

```text
[subject doing action] in [setting], [time of day], [camera framing/movement],
[lighting/mood]
```

```text
A samurai walking through a misty bamboo forest at dawn, low handheld tracking
shot from behind, soft golden backlight filtering through leaves
```

--------------------------------

### Search Fal.ai Documentation

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Search Fal.ai's documentation, guides, and API references using a query string. Useful for finding information on specific topics or features.

```bash
genmedia docs "how to use LoRA"
```

```bash
genmedia docs "webhook callbacks"
```

--------------------------------

### Kling v3 Multi-Prompt Sequence Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-prompting/references/kling.md

For Kling v3, when the schema supports multi-prompt sequences, structure each shot as a declarative line. Maintain cross-shot anchors like character and location.

```text
SHOT 1: wide establishing of the bamboo forest at dawn
SHOT 2: medium tracking shot of the samurai walking forward
SHOT 3: close-up of his hand on the sword hilt
```

--------------------------------

### Motion Anchor Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/storytelling.md

Maintains temporal continuity by ensuring a single continuous shot with no time jumps, starting where the previous shot ended.

```text
single continuous shot, starts where previous shot ended, no time jump
```

--------------------------------

### Generate Video with Start and End Frames

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/references/PATTERNS.md

This pattern generates a video using a prompt and specified start and end images. It requires nodes for the start frame, end frame, and the prompt.

```json
"node-video": {
  "depends": ["node-start-frame", "node-end-frame", "node-prompt"],
  "app": "fal-ai/kling-video/o1/image-to-video",
  "input": {
    "prompt": "$node-prompt.output",
    "image_url": "$node-start-frame.images.0.url",
    "tail_image_url": "$node-end-frame.images.0.url"
  }
}
```

--------------------------------

### Manage Installed Agent Skills

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Commands to install, update, list, and remove agent skills from the genmedia registry.

```bash
genmedia skills list
```

```bash
genmedia skills install genmedia
```

```bash
genmedia skills update
```

```bash
genmedia skills remove genmedia
```

--------------------------------

### Step 2: Perform Virtual Try-On

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/virtual-tryon.md

Execute the virtual try-on model using the cleaned garment image and person image. Specify the garment type and quality preference for optimal results.

```bash
URL_PERSON=$(genmedia upload ./person.jpg --json | jq -r '.url')

RES_TRYON=$(genmedia run fal-ai/fashn/tryon/v1.5 \
 --model_image "$URL_PERSON" \
 --garment_image "$URL_GARMENT_CLEAN" \
 --garment_type "top" \
 --quality "quality" \
 --json)
URL_TRYON=$(echo "$RES_TRYON" | jq -r '.image.url')
```

--------------------------------

### SaaS Device Mockup Prompt

Source: https://github.com/fal-ai-community/skills/blob/main/skills/commercial/references/examples.md

Create a professional mockup of a laptop displaying a SaaS dashboard. This prompt emphasizes a clean, productive workspace with specific lighting and depth of field.

```text
thin laptop on a clean walnut desk showing an abstract dashboard interface,
morning side light, shallow depth of field, organized workspace with notebook
and pen, calm productivity mood, wide landscape crop with negative space on
right for headline, no readable fake UI text
```

--------------------------------

### Smart Routing JSON Output Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Example of JSON output when using 'genmedia run' with smart routing, including the 'routed' block detailing the model selection.

```json
{
  "status": "completed",
  "endpoint_id": "fal-ai/flux/schnell",
  "routed": {
    "modality": "text-to-image",
    "source": "manifest",
    "from_prompt": "a cat on the moon"
  },
  "result": { ... }
}
```

--------------------------------

### Run FAL Site Upgrade Script

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-redesign/SKILL.md

Use this command when the user references a file path or a local URL. Ensure FAL_KEY is set.

```bash
node runtime/bin/fal-site.mjs upgrade
```

--------------------------------

### Product Anchor Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/storytelling.md

Maintains product consistency by specifying shape, color, packaging, logo placement, and material.

```text
same product shape, color, packaging, logo placement, and material as the
reference
```

--------------------------------

### Upload Product Image and Run Edit

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/product-shot.md

Uploads a product image and then runs an image editing process to create a hero shot. The `--json` flag outputs results in JSON format, and `jq` is used to extract the image URL.

```bash
URL_PRODUCT=$(genmedia upload ./product.png --json | jq -r '.url')

genmedia run fal-ai/nano-banana-pro/edit \
 --image_urls "$URL_PRODUCT" \
 --prompt "<structured product hero prompt>" \
 --download "./outputs/product-shot/{request_id}_{index}.{ext}" \
 --json
```

--------------------------------

### Hook Examples for UGC

Source: https://github.com/fal-ai-community/skills/blob/main/skills/ugc/references/formats.md

A collection of potential hooks to grab viewer attention. Remove any hook that implies a fact not provided by the user.

```text
"I did not expect [specific result] from [product category]."
```

```text
"If you are dealing with [problem], watch this part."
```

```text
"This is the smallest detail that changed [workflow/use case]."
```

```text
"Here is what I would check before buying [category]."
```

```text
"I used [product] for [scenario], and this was the useful part."
```

--------------------------------

### Creating Directories for Game Assets

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-gamedev/SKILL.md

Ensure the correct folder structure for game assets by creating directories using `mkdir -p` before downloading files. This applies even when only generating a subset of assets.

```bash
mkdir -p ./game-assets/<character-slug>/sprites
mkdir -p ./game-assets/<character-slug>/backgrounds
```

--------------------------------

### Discover Alternative Try-On Models

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/virtual-tryon.md

Search for alternative models related to virtual try-on or garment transfer, and access their documentation.

```bash
genmedia models "virtual tryon" --json
```

```bash
genmedia models "garment transfer" --json
```

```bash
genmedia docs "virtual tryon" --json
```

--------------------------------

### Character Anchor Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/storytelling.md

Ensures consistency in character appearance across shots by maintaining face, hair, wardrobe, age, posture, and style.

```text
same character face, hair, wardrobe, age, posture, and style as the approved
reference
```

--------------------------------

### Upload File and Run Model

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/SKILL.md

Upload a local file or remote URL to the fal.ai CDN, then use the returned URL to run a model. The output can be downloaded directly.

```bash
URL=$(genmedia upload ./photo.jpg --json | jq -r '.url')
genmedia run fal-ai/nano-banana-pro/edit \
 --image_urls "$URL" \
 --prompt "make the sky stormy" \
 --download "./out/{request_id}_{index}.{ext}" \
 --json
```

--------------------------------

### Discover Text-to-Video Models

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-models-catalog/references/text-to-video.md

Use the `genmedia` CLI to list available text-to-video models by category or to search for documentation. The `--json` flag ensures output is in a parseable format.

```bash
genmedia models --category text-to-video --limit 10 --json
```

```bash
genmedia docs "text to video" --json
```

--------------------------------

### Get Endpoint Schema

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/character-lipsync.md

Inspect the available parameters and schema for a specific endpoint before running it. This is useful for understanding required inputs and optional configurations.

```bash
genmedia schema veed/fabric-1.0 --json
```

```bash
genmedia schema fal-ai/sync-lipsync/v2 --json
```

--------------------------------

### Generate Audio with Multi-Prompt Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fan-cam/references/prompt-contract.md

This JSON object demonstrates how to generate audio for a fan cam scene using multiple prompts. Ensure 'generate_audio' is true and do not combine 'multi_prompt' with 'end_image_url'. For native audio, specify it as an off-screen voiceover and include 'no lip sync'.

```json
{
  "duration": "6",
  "generate_audio": true,
  "shot_type": "customize",
  "cfg_scale": 0.32,
  "start_image_url": "<approved 16:9 spectator frame>",
  "multi_prompt": [
    {
      "duration": "3",
      "prompt": "Use the start image as the same live NBA timeout crowd shot. The featured spectator is silent and watches the missed t-shirt cannon moment with a small amused reaction. No visible person speaks."
    },
    {
      "duration": "3",
      "prompt": "Keep the camera on the same spectator and crowd. An off-camera arena PA or crowd-mic voice from the venue says, \"that was close!\" It is not the spectator speaking; no lip sync, no mouth movement matching the words."
    }
  ],
  "negative_prompt": "featured spectator speaking, lip sync, lips moving to narration, mouth forming words, on-camera subject says the line, dialogue from visible fan, visible fan voice, face morphing, unstable scoreboard, wrong sport, camera leaving crowd, end image url"
}

```

--------------------------------

### Generate a new site with generate command

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-redesign/SKILL.md

Use the `generate` command for greenfield site creation, transforming a freeform context into a visual mockup and then a single-file HTML site. Use `--variants` to explore multiple initial concepts.

```bash
node runtime/bin/fal-site.mjs generate --context "<freeform context>" [--variants 4] [--out <dir>]
```

--------------------------------

### FAL Workflow Troubleshooting: LLM Vision Error

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/SKILL.md

Guides users to fix the 'LLM Vision Error' by switching to 'openrouter/router/vision' when analyzing images.

```text
Error: image_urls provided but using text-only router

**Solution:** Switch to `openrouter/router/vision` when analyzing images.
```

--------------------------------

### Search Image-to-Video Documentation

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-models-catalog/references/image-to-video.md

Use the `genmedia docs` command to search for documentation related to image-to-video generation. The `--json` flag provides structured output.

```bash
genmedia docs "image to video" --json
```

--------------------------------

### Smart Routing with Genmedia Run

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Execute models using smart routing by providing a prompt. The CLI automatically determines the modality and routes to a default endpoint. Use --json for JSON output.

```bash
genmedia run "a cat on the moon" --json
```

```bash
genmedia run "a 5-second clip of a robot dancing" --async --json
```

```bash
genmedia run "narrate this paragraph in a calm voice" --json
```

```bash
genmedia run --prompt "a cat on the moon" --json # equivalent: --prompt without positional
```

--------------------------------

### Manifest Node Example

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/references/node-rules.md

A compact manifest structure to be returned at the end of a workflow. It should contain factual information about the executed node, its inputs, outputs, and status.

```json
{
  "node_id": "shot_03_i2v",
  "role": "image_to_video",
  "endpoint_id": "selected endpoint",
  "request_id": "fal request id",
  "inputs": ["source url or local path"],
  "outputs": ["result media url"],
  "downloaded_files": ["local file path"],
  "status": "accepted | retried | rejected",
  "notes": "short defect or continuity note"
}
```

--------------------------------

### Sneaker Hover Hero Shot Prompt

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/product-shot.md

Example prompt for generating a hero shot of a sneaker in a suspended state, detailing its appearance, environment, and lighting.

```text
single white running sneaker preserved exactly from the reference, blue
outsole accents intact, three-quarter side view, suspended just above clean
concrete, sharp shadow under sole, crisp studio flash, slight motion dust
behind heel, athletic launch campaign style, 4:5 crop with safe space upper
right for headline
```

--------------------------------

### Skincare Bottle Hero Shot Prompt

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/product-shot.md

Example prompt for generating a hero shot of a skincare bottle, specifying material, surface, lighting, and background.

```text
amber glass skincare serum bottle preserved exactly from the reference, matte
black dropper, standing on pale travertine, soft diffused key light from upper
left, thin rim light on glass edges, subtle water condensation, warm beige
background, square crop, premium beauty product photography, no added text,
no extra bottles
```

--------------------------------

### Discover Image-to-Image Models

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-models-catalog/references/image-to-image.md

Use the `genmedia` CLI to discover available image-to-image models, search for image editing models, or find documentation related to image editing.

```bash
genmedia models --category image-to-image --limit 10 --json
```

```bash
genmedia models "image edit" --json
```

```bash
genmedia docs "image editing" --json
```

--------------------------------

### Run Video Generation Asynchronously with Genmedia

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/character-design.md

Initiate asynchronous video generation using Genmedia. This command requires an endpoint ID, a prompt detailing the character and action, and optionally a reference frame URL. The `--async` flag enables background processing, and `--json` ensures structured output.

```bash
genmedia run <endpoint_id> \
 --prompt "<anchor + shot action>" \
 --image_url "<approved character frame if supported>" \
 --async \
 --json
```

--------------------------------

### Get Kling v3 Schema

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-prompting/references/kling.md

Use the `genmedia schema` command to inspect the available fields for multi-prompting and other features specific to Kling v3 endpoints.

```bash
genmedia schema fal-ai/kling-video/v3/pro/text-to-video --json
```

--------------------------------

### Problem-Solution Demo Structure

Source: https://github.com/fal-ai-community/skills/blob/main/skills/ugc/references/formats.md

Ideal for products with visible use cases. Follow this sequence: introduce the problem, show the product, demonstrate a clean action, and reveal the result.

```text
Start with the frustrating moment. Show the product entering the scene. Show
one clean action. Show the result. End on the product or next step.
```

--------------------------------

### Get Text-to-3D Schema

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-models-catalog/references/text-to-3d.md

Retrieve the schema for a specific Text-to-3D model to understand its available parameters. This is useful for discovering model-specific options like `art_style`.

```bash
genmedia schema fal-ai/meshy/v6/text-to-3d --json
```

--------------------------------

### MMAudio Video to Audio Extraction

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/references/MODELS.md

Configuration for MMAudio to extract audio from a video. Requires a video URL and an optional prompt to guide audio extraction.

```json
{
  "app": "fal-ai/mmaudio-v2",
  "input": {
    "video_url": "$node-video.video.url",
    "prompt": "Ambient nature sounds"
  }
}
```

--------------------------------

### Manage Genmedia Version and Updates

Source: https://github.com/fal-ai-community/skills/blob/main/skills/genmedia/references/full-reference.md

Check the current version, download and install updates, or perform update checks. Automatic background updates are enabled by default.

```bash
genmedia version # current version + any pending update
```

```bash
genmedia update # download and swap in latest
```

```bash
genmedia update --check # check only, no download
```

```bash
genmedia update --force # reinstall even if already on latest
```

--------------------------------

### Run Photorealism Generation with Genmedia

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/realism.md

Execute a photorealistic image generation using a structured prompt and specified parameters. Ensure to replace '<structured prompt>' with your actual prompt.

```bash
# 3. Run with structured prompt
genmedia run openai/gpt-image-2 \
  --prompt "<structured prompt>" \
  --image_size landscape_4_3 \
  --quality high \
  --num_images 1 \
  --output_format png \
  --download "./outputs/realism/{request_id}_{index}.{ext}" \
  --json
```

--------------------------------

### Extract Frame and Generate Video

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-workflow/references/PATTERNS.md

This pattern extracts a frame from a video and uses it as the starting frame for a new video generation. It depends on the extracted frame and a prompt.

```json
"node-extract": {
  "depends": ["node-video-1"],
  "app": "fal-ai/ffmpeg-api/extract-frame",
  "input": {
    "video_url": "$node-video-1.video.url",
    "frame_type": "last"
  }
},
"node-video-2": {
  "depends": ["node-extract", "node-prompt-2"],
  "app": "fal-ai/kling-video/o1/image-to-video",
  "input": {
    "prompt": "$node-prompt-2.output",
    "image_url": "$node-extract.frame.url"
  }
}
```

--------------------------------

### Coffee Mug Background Swap Prompt

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-recipes/references/product-shot.md

Example prompt for replacing the background of a coffee mug image with a kitchen scene, specifying lighting and desired style.

```text
preserve the uploaded ceramic coffee mug exactly, replace background with a
bright Scandinavian kitchen counter, soft morning window light from left,
realistic contact shadow and reflection on counter, warm lifestyle product
photography, no new logo, no deformation, no text
```

--------------------------------

### Inspect Endpoint Schema and Pricing

Source: https://github.com/fal-ai-community/skills/blob/main/skills/character-design/SKILL.md

Before running an endpoint, inspect its schema to understand supported parameters and check pricing information.

```bash
genmedia schema <endpoint_id> --json
```

```bash
genmedia pricing <endpoint_id> --json
```

--------------------------------

### Get Schema for OCR Endpoint

Source: https://github.com/fal-ai-community/skills/blob/main/skills/fal-models-catalog/references/image-to-text.md

Use this command to retrieve the schema for the GOT OCR 2.0 endpoint. This helps in understanding the expected input and output formats.

```bash
genmedia schema fal-ai/got-ocr/v2 --json
```