### Install Gen-AI Skills using Setup Script

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

A universal fallback method that clones the repository and symlinks skills. Can override the host agent with `--host <agent>`.

```bash
git clone --depth 1 https://github.com/PicsArt/gen-ai-skills.git
cd gen-ai-skills
./setup
```

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

### Setup Local Environment

Source: https://github.com/picsart/gen-ai-skills/blob/main/CONTRIBUTING.md

Run the setup script locally to prepare your environment and verify the new skill loads in your agent.

```bash
./setup
```

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

### Parallel Stress Test Setup

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/references/BATCH.md

Example of creating a temporary manifest file for a parallel stress test with multiple job variants for the same model. Ensure unique IDs for each job.

```bash
cat > /tmp/nb2.json <<'JSON'
{
  "defaults": { "aspect_ratio": "1:1" },
  "jobs": [
    { "id": "01-neon-city",     "model": "gemini-3.1-flash-image", "prompt": "neon cyberpunk street at night, rain reflections" },
    { "id": "02-koi-pond",      "model": "gemini-3.1-flash-image", "prompt": "overhead koi pond with lily pads" },
    { "id": "03-astronaut-cat", "model": "gemini-3.1-flash-image", "prompt": "astronaut cat floating in space, earth in visor" }
  ]
}
JSON
# Note: to run 20 jobs, repeat the same object shape for ids 04..20 with distinct prompts.
# JSON does not support comments, so do not paste a "/* ... */" placeholder into the manifest.

gen-ai batch run /tmp/nb2.json -c 20 -o ./batch-output/nb2-parallel
```

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

### Install Gen-AI Skills using gh skill install

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Uses the GitHub CLI extension (v2.90+). Installs all skills from the repository.

```bash
gh skill install PicsArt/gen-ai-skills
```

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

### Install and Authenticate Gen AI CLI

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-persona-creation/SKILL.md

Verify gen-ai installation and authentication. If authentication fails, use `gen-ai login` or set environment variables.

```bash
gen-ai whoami            # auth + gen-ai install + Node v22+ check
command -v curl          # ships with macOS / Linux / Git-Bash
```

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

### Install and Use Gen AI CLI

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-client-handoff/SKILL.md

Steps to install the Gen AI CLI and generate an asset using a prompt from the library. Ensure you are logged in before generating.

```bash
1. Install CLI: curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash
2. gen-ai login
3. Pick a prompt from prompts/prompt-library.json
4. gen-ai generate -m <pinned-model-id> -p "<prompt>"
```

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

### Install Picsart Gen-AI CLI

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/product-photo-studio/SKILL.md

Installs the Picsart Gen-AI CLI using a recommended script.

```bash
curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash
```

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

### Install gen-ai-skills CLI

Source: https://github.com/picsart/gen-ai-skills/blob/main/README.md

Use this command to add the PicsArt/gen-ai-skills package to your project. Refer to INSTALL.md for alternative installation methods.

```bash
npx skills add PicsArt/gen-ai-skills
```

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

### Update Gen-AI Skills using Setup Script

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Commands to update skills installed via the setup script by pulling the latest changes and re-running the script.

```bash
cd gen-ai-skills && git pull && ./setup
```

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

### Install Picsart Gen AI CLI

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/SKILL.md

Use this command to install the CLI via curl. Alternatively, install via npm if Node.js is available.

```bash
# Signed binary (no Node required) — recommended
curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash

# Or via npm (requires Node 22+)
npm install -g @picsart/gen-ai

gen-ai login     # OAuth browser flow
gen-ai whoami    # verify
```

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

### Install and Authenticate Picsart Gen-AI CLI

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/multi-channel-bundle/SKILL.md

Install the Picsart Gen-AI CLI and authenticate your account to use the tool. This is a prerequisite for running any gen-ai commands.

```bash
curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash
gen-ai login
gen-ai whoami
```

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

### Multi-Brand Pack Workflow Steps

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-multi-brand-pack/SKILL.md

Follow these steps to execute the multi-brand pack generation process. Ensure all prerequisites are met before starting.

```bash
1. VERIFY BRANDS  → every client in scope has clients/<slug>/brand.md + brand-system.json
2. SCAFFOLD       → build per-client manifest fragments, merge into one master manifest
3. ESTIMATE       → gen-ai batch run <manifest.json> --dry-run + estimate; break into smaller batches if > $20
4. RUN            → `gen-ai batch run <manifest.json>` with client names in job IDs and a resume-ready output dir
5. VERIFY         → per-client output folders populated; cross-check any NULLs in results.json
6. AUDIT          → export per-client spend + asset list for billing
7. DELIVER        → folder per client, approved files uploaded or linked in the retainer channel
```

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

### Install a Single Skill with Hermes Agent

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Installs a specific skill from the repository using the Hermes Agent.

```bash
hermes skills install PicsArt/gen-ai-skills/skills/gen-ai-use
```

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

### Local Image Resizing Tool Example

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/dev-app-assets/SKILL.md

Use local tools like sips, ImageMagick, or sharp for resizing assets below the master size to prevent AI-driven drift.

```bash
sips -Z 1024 icon-master.png --out icon-master-1024.png
```

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

### Manifest Schema Example

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/references/BATCH.md

An example of a manifest file in JSON format, defining default parameters and a list of jobs with specific models and prompts. Image-edit models can also accept an 'image' parameter.

```json
{
  "defaults": { "aspect_ratio": "1:1" },
  "jobs": [
    { "id": "hero",   "model": "flux-2-pro",             "prompt": "sunset over mountains", "aspect_ratio": "16:9" },
    { "id": "cat",    "model": "gemini-3.1-flash-image", "prompt": "a cat in space" },
    { "id": "remix",  "model": "flux-kontext-pro",       "prompt": "make it watercolor",    "image": "./src.jpg" }
  ]
}
```

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

### Skill Frontmatter Example

Source: https://github.com/picsart/gen-ai-skills/blob/main/CONTRIBUTING.md

The frontmatter for a SKILL.md file must include all specified fields. CI will fail if any are missing.

```markdown
---
name: picsart-upscale
description: One-sentence description ending with a period.
version: 1.0.0
author: Picsart
license: MIT
platforms: [macos, linux]
metadata:
  hermes:
    category: creative
    tags: [picsart, upscale, image-quality]
---

# Picsart Upscale

Two-to-three-sentence intro stating what the skill does and what it isn't for.

## When to Use
## Prerequisites
## How to Run
## Quick Reference
## Procedure
## Pitfalls
## Verification

```

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

### Interactive Generation

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/SKILL.md

Launches an interactive wizard for guided generation tasks.

```bash
gen-ai generate
```

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

### Brand System Configuration Example

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-brand-scoping/SKILL.md

This JSON configuration defines parameters for brand system generation, including client specifics, default models, aspect ratios, and job variants.

```json
{
  "client_slug": "acme-fintech",
  "brand_system": "clients/acme-fintech/brand-system.json",
  "defaults": { "model": "recraftv4", "aspectRatio": "16:9" },
  "variants": {
    "direction": ["editorial", "bold", "minimal", "playful", "cinematic"]
  },
  "jobs": [
    { "id": "{direction}", "prompt": "<client descriptor> — {direction} direction" }
  ]
}
```

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

### Client Handoff Manifest Example

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-client-handoff/SKILL.md

A JSON manifest file that provides a comprehensive overview of the delivered assets and their associated metadata. Include this as `docs/handoff-manifest.json`.

```json
{
  "handoff_kind": "client-final",
  "client_slug": "acme-fintech",
  "engagement": "2025-Q3 to 2026-Q2 retainer",
  "delivered_at": "2026-04-22",
  "brand_system": "docs/brand-system.json",
  "brand_rules": "docs/brand.md",
  "pinned_models": "docs/model-pins.json",
  "assets_count": 247,
  "prompts_count": 54,
  "rights_status": "all generated, no licensed stock",
  "regeneration_supported": true,
  "support_window_days": 30
}
```

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

### Model Information Commands

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/references/FLAGS.md

Commands to list, get details, and compare generative AI models.

```bash
gen-ai models                       # List all models (filter --mode, --provider, --input-type, --params)
```

```bash
gen-ai models:info <id>             # (or `models info <id>`) — capabilities, inputs, defaults
```

```bash
gen-ai models:compare <a> <b>       # Side-by-side comparison
```

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

### Example Script Presentation Format

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-explainer/references/script-director.md

Illustrates the structured format for presenting a script in a chat-friendly manner, including sections, word counts, and voice directions.

```text
**Section 1 — Hook (0-5s, 18 words):** "Your database searches every single row.
Every. Single. One. What if it didn't have to?"
  Voice: emphasize 'every single row' with measured pacing.

**Section 2 — Setup (5-15s, 28 words):** ...

**Section 3 — Build (15-28s, 35 words):** ...
...

Total: 4 sections, 142 words, ~57s at 150 wpm. Budget for 60s = 130-150 words. ✓
```

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

### Example Scene Plan

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-explainer/references/scene-plan-director.md

This JSON object defines a scene with an image prompt, motion hint, and duration. It serves as a template for creating detailed scene descriptions for AI generation.

```json
{
  "id": "s3",
  "image_prompt": "Isometric cutaway of an espresso machine boiler chamber, copper pipes carrying water through a heating element, pressure gauge visible in the foreground reading '9 BAR', soft pastel medical-illustration palette matching scene 1, 16:9, centered, eye-level perspective",
  "motion_hint": "Water flows visibly through the pipes from left to right while the pressure gauge needle slowly rises from 1 to 9 over 3 seconds, then holds. Slow zoom on the gauge in the final second.",
  "duration_s": 12
}
```

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

### Generate Image and Animate to Video

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/references/EXAMPLES.md

Start by generating an initial image with a specified model and prompt, then use the generated image to create a video with a different model and prompt.

```bash
# Generate initial image
gen-ai generate --model flux-2-pro --prompt "cyberpunk cityscape at night"

# Turn it into a video
gen-ai generate --model kling-v3-pro --image ~/Downloads/result.png \
  --prompt "camera slowly pans across the city"
```

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

### Generate Pilot Scene

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/product-photo-studio/references/modes/compose.md

Use this command to generate an initial pilot scene for a product. Iterate on the prompt until scale, shadow, and product fidelity are satisfactory. Ensure the input image and prompt details align with desired output.

```bash
gen-ai generate -m runway-gen4-ref -i ./hero/SKU-042.png \
  -p "on a white marble kitchen counter, morning light from the left, soft cast shadow, blurred background of a bright airy kitchen" \
  --ar 4:5 --download ./pilot
```

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

### Initialize Handoff Directory Structure

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-client-handoff/SKILL.md

Creates a structured directory for handoff assets, prompts, source files, and documentation. It copies relevant files from the client's project.

```bash
CLIENT="acme-fintech"
HANDOFF="handoff/$CLIENT-$(date +%Y-%m-%d)"
mkdir -p "$HANDOFF"/{assets,prompts,source,docs}

find "clients/$CLIENT" -name "results.json" -exec cp {} "$HANDOFF/source/" \;
cp -r "clients/$CLIENT/deliverables/"* "$HANDOFF/assets/"
node scripts/extract-prompts.js "$HANDOFF/source/" > "$HANDOFF/prompts/prompt-library.json"
```

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

### Update Gen-AI Skills with gh skill install

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Command to update skills installed using the GitHub CLI extension.

```bash
gh skill update PicsArt/gen-ai-skills
```

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

### Agency Pitch Mockup Workflow Steps

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-pitch-mockups/SKILL.md

This list outlines the sequential steps involved in generating and delivering agency pitch mockups. Follow these steps for a consistent and efficient workflow.

```text
1. LOAD BRAND    → read clients/<slug>/brand.md + brand-system.json
2. PLAN          → table of every asset with prompt + aspect + model
3. ESTIMATE      → gen-ai pricing on the full manifest (target < $5)
4. DRAFT PASS    → run at draft quality (gemini-3.1-flash-image / recraftv4) for fast iteration
5. CLIENT REVIEW → internal pitch-team review on the draft set
6. FINALIZE      → upgrade approved concepts to flagship models
7. WATERMARK     → add subtle "MOCKUP" footer where required
8. DELIVER       → organized folder + approved upload/link + deck-ready filenames
```

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

### Inspect Sample Output

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/product-photo-studio/SKILL.md

Use this command to perform a spot-check on the dimensions of a single generated output file. Replace `<run>` with your specific run identifier and `<sample>.jpg` with the filename of the asset you want to inspect.

```bash
gen-ai inspect outputs/<run>/<sample>.jpg
```

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

### Upload Explainer Video to Drive

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-explainer/SKILL.md

After rendering and finalizing the explainer video, use this command to upload it to Google Drive. Provide the video file path and a desired title for the upload.

```bash
gen-ai upload-to-drive <slug>/explainer.mp4 --name "<title>"
```

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

### Pronunciation Guides for TTS

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-explainer/references/script-director.md

This JSON array provides pronunciation guides for technical terms, acronyms, and brand names to ensure correct pronunciation by Text-to-Speech engines. Include any term that might not be pronounced correctly by default.

```json
[
  { "word": "FAISS", "phonetic": "FACE" },
  { "word": "Qdrant", "phonetic": "kuh-DRANT" },
  { "word": "GPT", "phonetic": "gee pee tee" }
]
```

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

### Client Handoff Process Steps

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-client-handoff/SKILL.md

Outlines the sequential steps for preparing and delivering AI-generated assets to clients. Follow these steps to ensure a complete and reproducible handoff.

```text
1. INVENTORY   → pull every asset + results.json from clients/<slug>/ across the engagement
2. FILTER      → exclude drafts, rejected variants, internal-only WIP
3. STRIP       → white-label: remove agency tags, internal prompt notes, competitor refs
4. PIN MODELS  → lock every prompt's model to an exact ID + version (not "latest")
5. DOCUMENT    → generate README, RIGHTS.md, CHANGELOG, prompt library
6. PACKAGE     → zip with consistent folder structure + naming
7. VERIFY      → extract and cold-test: can a fresh machine regenerate one asset?
8. DELIVER     → upload to client destination, share link, brief the receiving team
```

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

### Get Logo Generation Result

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/picsart-api/references/genai-api.md

Polls the status and retrieves the result of a logo generation job.

```APIDOC
## `genai-generate-logo-getresult`

### Description
Poll the logo generation result.

### Method
GET

### Endpoint
`/logo/inferences/{inference_id}`

### Parameters
#### Path Parameters
- **inference_id** (string, required) - The ID of the inference job.

### Response
#### Success Response (200)
- **status** (string) - The status of the logo generation.
- **data** (array) - Array of logo objects.
  - **id** (string) - Unique identifier for the logo.
  - **url** (string) - URL to the generated logo.
  - **status** (string) - Status of the individual logo.

#### Processing Response (202)
Indicates that the job is still in progress.

### Response Example (200)
```json
{
  "status": "completed",
  "data": [
    {
      "id": "logo-abc",
      "url": "https://picsart.com/logo.png",
      "status": "completed"
    }
  ]
}
```

### Response Example (202)
```json
{
  "status": "processing"
}
```
```

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

### Pricing and Credit Commands

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/references/FLAGS.md

Commands to check model pricing and user credit balance.

```bash
gen-ai pricing <model-id>           # Credit cost for a model
```

```bash
gen-ai credits                      # Show current credit balance
```

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

### Get Export Result

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/picsart-api/references/variable-data-api.md

Poll for the completion status and results of a variable data export transaction.

```APIDOC
## Get Export Result

### Description
Poll for the completion status and results of a variable data export transaction.

### Method
GET

### Endpoint
/export-variable-data-content/getresult

### Parameters
#### Query Parameters
- **transaction_id** (string) - Required - The ID of the export transaction.

### Response
#### Success Response (200)
- **data** (array) - An array of results, where each element contains 'urls' (array of strings) for the generated content.

#### Pending Response (202)
Indicates the job is still processing. Retry later.

### Response Example (Success)
```json
{
  "data": [
    {
      "urls": ["https://cdn.example.com/output1.pdf"]
    },
    {
      "urls": ["https://cdn.example.com/output2.pdf"]
    }
  ]
}
```
```

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

### Lifestyle Product Scene Mockups (SaaS/Hardware)

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/product-photo-studio/references/modes/mockups.md

Configure jobs for generating lifestyle mockups of products like phones or laptops. Define defaults for model and aspect ratio, then specify prompts and image URLs for each job.

```json
{
  "defaults": { "model": "flux-2-pro", "aspectRatio": "16:9" },
  "jobs": [
    { "id": "phone-in-hand", "prompt": "modern app UI on an iPhone 15 Pro held in one hand, cafe background, shallow DoF", "imageUrls": ["./screen.png"] },
    { "id": "laptop-desk",   "prompt": "same app on a MacBook Pro at a minimal wooden desk, morning light, plant in frame" },
    { "id": "billboard",     "prompt": "same product packaging on an urban billboard at dusk, cinematic wide shot" }
  ]
}
```

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

### video-trim

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/picsart-api/references/video-api.md

Trims a video based on specified start and end timestamps (in milliseconds). This is an asynchronous operation.

```APIDOC
## video-trim

### Description
Trim by start/end ms.

### Key Inputs
`video_url`, `start`, `end`, ExportParameters

### Type
Async (`video-getresult`)
```

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

### Update Gen-AI Skill with Hermes Agent

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Command to update a specific skill installed through Hermes Agent.

```bash
hermes skills update gen-ai-use
```

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

### Get Replay Export Result

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/picsart-api/references/variable-data-api.md

Poll for the completion status and retrieve the URL for a one-off Replay export.

```APIDOC
## Get Replay Export Result

### Description
Poll for the completion status and retrieve the URL for a one-off Replay export.

### Method
GET

### Endpoint
/export-replay/getresult

### Parameters
#### Query Parameters
- **transaction_id** (string) - Required - The ID of the export transaction.

### Response
#### Success Response (200)
- **data.url** (string) - The URL of the generated export.

#### Pending Response (202)
Indicates the job is still processing. Retry later.

### Response Example (Success)
```json
{
  "data": {
    "url": "https://cdn.example.com/output.pdf"
  }
}
```
```

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

### Update Gen-AI Skills with npx skills

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Command to re-run for updating skills installed via `npx skills`.

```bash
re-run `npx skills add PicsArt/gen-ai-skills`
```

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

### Get Sticker Generation Result

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/picsart-api/references/genai-api.md

Polls the status and retrieves the result of a sticker generation job (standard or laser-engraving).

```APIDOC
## `genai-text2sticker-getresult`

### Description
Poll result for both standard and laser-engraving sticker submissions.

### Method
GET

### Endpoint
`/text2sticker/inferences/{inference_id}`

### Parameters
#### Path Parameters
- **inference_id** (string, required) - The ID of the inference job.

### Response
#### Success Response (200)
- **status** (string) - The status of the generated sticker.
- **data** (array) - Array of sticker objects.
  - **id** (string) - Unique identifier for the sticker.
  - **url** (string) - URL to the generated sticker.
  - **status** (string) - Status of the individual sticker.

#### Processing Response (202)
Indicates that the job is still in progress.

### Response Example (200)
```json
{
  "status": "completed",
  "data": [
    {
      "id": "sticker-123",
      "url": "https://picsart.com/sticker.png",
      "status": "completed"
    }
  ]
}
```

### Response Example (202)
```json
{
  "status": "processing"
}
```
```

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

### GET /export/replay/{transaction_id}

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/picsart-api/references/variable-data-api.md

Retrieves the finished result, such as a PDF URL, for a previously initiated asynchronous export operation.

```APIDOC
## GET /export/replay/{transaction_id}

### Description
Retrieve the finished result (e.g. PDF URL) for an earlier `vd-export-replay` call.

### Method
GET

### Endpoint
/export/replay/{transaction_id}

### Parameters
#### Path Parameters
- **transaction_id** (string, required) — The ID returned by `vd-export-replay`.

#### Query Parameters
None

#### Request Body
None

### Request Example
```json
{
  "example": "GET /export/replay/some-transaction-id"
}
```

### Response
#### Success Response (202 - Still Processing)
- **transaction_id** (string) - The ID for tracking the asynchronous operation.
- **status** (string) - The current status of the operation.

#### Success Response (200 - Completed)
- **data** (object) - Contains the result details.
  - **id** (string) - The unique identifier for the generated file.
  - **url** (string URI) - The URL pointing to the generated file.
- **status** (string) - The final status of the operation.

#### Response Example (Processing)
```json
{
  "transaction_id": "some-transaction-id",
  "status": "processing"
}
```

#### Response Example (Completed)
```json
{
  "data": {
    "id": "generated-file-id",
    "url": "https://example.com/path/to/your/file.pdf"
  },
  "status": "completed"
}
```

**Async?** This is the polling endpoint paired with `vd-export-replay`.

**Tips:**
- Poll with a small backoff (e.g. 1–3s) until you get the `200` shape with a `url`.
- The `url` is the downloadable result; persist it promptly if you need it long-term.
```

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

### Set Up CI Authentication

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/text-to-visual/references/modes/og.md

Configure environment variables for Picsart API authentication in a CI/CD environment. This is necessary for headless runs of the gen-ai tools.

```bash
export PICSART_ACCESS_TOKEN="${{ secrets.PICSART_TOKEN }}"
export PICSART_USER_ID="${{ secrets.PICSART_USER_ID }}"
```

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

### Estimate Background Swap Cost

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/dev-screenshot-beautifier/SKILL.md

Use these commands to get information about the recraftv3-replace-bg model and its pricing before performing a background swap.

```bash
gen-ai models info recraftv3-replace-bg
gen-ai pricing recraftv3-replace-bg
```

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

### Generate Image-to-Video

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/SKILL.md

This command generates video from an input image, using the image as the starting frame. Specify the duration with -d.

```bash
gen-ai generate -m veo-3.1 -i ~/photo.jpg -p "camera zooms in" -d 5
```

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

### Run Batch Generation with Manifest

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/multi-channel-bundle/SKILL.md

Execute a batch generation process using a manifest file to create a bundle of assets. This command estimates pricing before committing to generation.

```bash
# Bundle generation is always a batch run
gen-ai batch --manifest manifest.json --output bundle/

# Estimate before committing
gen-ai pricing --manifest manifest.json
```

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

### Configuration Management Commands

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/references/FLAGS.md

Commands to get, set, list, and unset configuration settings for the gen-ai CLI. Configuration persists to `~/.gen-ai/config.json`.

```bash
gen-ai config:get <key>          # Get a setting
```

```bash
gen-ai config:set <key> <value>  # Set a preference
```

```bash
gen-ai config:list               # Show all settings
```

```bash
gen-ai config:unset <key>        # Remove a preference
```

```bash
gen-ai config:keys               # List valid config keys
```

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

### Get Model Information

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/SKILL.md

Retrieve detailed information about a specific model, including its capabilities, input requirements, and supported aspect ratios.

```bash
gen-ai models info <id>             # capabilities, inputs, aspect ratios
```

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

### Check Pricing and Balance

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/gen-ai-use/references/TROUBLESHOOTING.md

Use this command to check your credit balance if you encounter an 'Credit insufficient' error.

```bash
gen-ai pricing <model>
```

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

### Run Batch Job - Sample Catalog

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/product-photo-studio/references/modes/reshoot.md

Executes a sample batch run for a catalog with specified concurrency and output directory.

```bash
gen-ai batch run catalog.sample.json --concurrency 3 \
  --output ./runs/2026-04-spring-sample
```

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

### Update Gen-AI Skills via Claude Code Marketplace

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Command to update the plugin for skills installed via the Claude Code marketplace.

```bash
/plugin update picsart@picsart
```

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

### Run Campaign Bundle in Parallel

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/multi-channel-bundle/references/modes/campaign.md

Generate the full campaign kit by running the batch job. Use the approved hero image as a reference and set concurrency for optimal performance.

```bash
gen-ai batch run campaign.json -c 4 -o ./out
```

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

### Agency Pitch Mockup Configuration Example

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-pitch-mockups/SKILL.md

This JSON object defines the configuration parameters for generating pitch mockups. It includes client details, concept, phase, brand system path, default model, asset variants, and a mockup footer.

```json
{
  "client_slug": "acme-fintech",
  "concept": "waits-for-you",
  "phase": "draft",
  "brand_system": "clients/acme-fintech/brand-system.json",
  "defaults": {
    "model": "recraftv4"
  },
  "variants": {
    "asset": ["hero", "tile-01", "tile-02", "tile-03", "quote-01", "quote-02"]
  },
  "mockup_footer": "MOCKUP — For internal review only"
}
```

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

### Install Gen-AI Skills via Claude Code Marketplace

Source: https://github.com/picsart/gen-ai-skills/blob/main/INSTALL.md

Specific to Claude Code. This method pulls the plugin manifest and registers the skills.

```bash
/plugin marketplace add PicsArt/gen-ai-skills
```

```bash
/plugin install picsart@picsart
```

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

### Create and Run Multi-Brand Manifest

Source: https://github.com/picsart/gen-ai-skills/blob/main/skills/agency-multi-brand-pack/SKILL.md

This section demonstrates how to create a JSON manifest file for batch AI generation jobs, specifying default and per-job configurations. It includes commands for a dry run and an actual execution with specified concurrency and output directory.

```bash
cat > /tmp/multi-brand-$WEEK.json <<'EOF'
{
  "defaults": { "model": "recraftv4", "aspectRatio": "1:1" },
  "jobs": [
    { "id": "acme-fintech/hero",    "prompt": "acme-fintech Q2 hero — editorial, restrained", "aspectRatio": "16:9" },
    { "id": "acme-fintech/og",      "prompt": "acme-fintech OG — headline: Q2 launch", "aspectRatio": "1200x630", "model": "ideogram-v3" },
    { "id": "zest-retail/hero",     "prompt": "zest-retail Q2 hero — playful, bold color", "aspectRatio": "16:9" },
    { "id": "nova-travel/hero",     "prompt": "nova-travel Q2 hero — cinematic wide", "aspectRatio": "16:9", "model": "flux-2-pro" }
  ]
}
EOF

gen-ai batch run /tmp/multi-brand-$WEEK.json --dry-run
gen-ai batch run /tmp/multi-brand-$WEEK.json -c 4 -o "out/retainer-$WEEK"
```