npx skills add https://github.com/fal-ai-community/skills --skill fal-generateCLI を使用してこのスキルをインストールし、ワークスペースで SKILL.md ワークフローの使用を開始します。
Agent skills for fal.ai — ready-to-use bash scripts that let AI agents generate images, videos, audio, 3D models, and more.
Compatible with Claude.ai Projects, Claude Code, and any agent platform that supports the community skills format.
| Skill | Description | Scripts |
|---|---|---|
| fal-redesign | Redesign a coded website end-to-end — screenshot, mockup, build-spec, HTML implementation, demo video | upgrade.sh, describe.sh, iterate.sh, generate.sh |
| fal-generate | Generate images and videos with queue support | generate.sh, upload.sh, search-models.sh, get-schema.sh |
| fal-image-edit | Edit images — style transfer, object removal, background change, inpainting | edit-image.sh |
| fal-audio | Text-to-speech and speech-to-text | text-to-speech.sh, speech-to-text.sh |
| fal-upscale | Upscale and enhance image resolution | upscale.sh |
| Skill | Description | Scripts |
|---|---|---|
| fal-kling-o3 | Kling O3 — highest quality photorealistic images and video editing | kling-generate.sh, kling-video.sh |
| fal-realtime | Real-time/streaming generation (~0.3s per image) | realtime.sh |
| Skill | Description | Scripts |
|---|---|---|
| fal-video-edit | Edit videos — remix, upscale, remove background, add audio | edit-video.sh, video-audio.sh |
| fal-lip-sync | Talking head, lip sync, live portrait | talking-head.sh, lip-sync.sh |
| Skill | Description | Scripts |
|---|---|---|
| fal-3d | Text/image to 3D model generation | generate-3d.sh |
| fal-vision | Image analysis — segment, detect, OCR, describe | analyze.sh |
| fal-restore | Restore image quality — deblur, denoise, fix faces | restore.sh |
| fal-tryon | Virtual clothing try-on | tryon.sh |
| fal-train | Train custom LoRA models | train.sh |
| Skill | Description | Scripts |
|---|---|---|
| fal-platform | Pricing, usage tracking, cost estimation, API key setup | pricing.sh, usage.sh, estimate-cost.sh, setup.sh, requests.sh |
| fal-workflow | Create multi-step model pipelines | create-workflow.sh |
Sign up at fal.ai/dashboard/keys.
export FAL_KEY=your_key_here
Or use the built-in setup:
bash scripts/generate.sh --add-fal-key
Upload skills to a Claude.ai Project:
claude.ai/ directory*.fal.ai is in your allowed domains under claude.ai/settings/capabilitiesEvery skill is a standalone bash script. Call them directly:
# Generate an image
bash skills/claude.ai/fal-generate/scripts/generate.sh \
--prompt "A serene mountain landscape" \
--model "fal-ai/nano-banana-pro"
# Generate a video
bash skills/claude.ai/fal-generate/scripts/generate.sh \
--prompt "Ocean waves crashing on rocks" \
--model "fal-ai/veo3.1" \
--async
# Edit an image
bash skills/claude.ai/fal-image-edit/scripts/edit-image.sh \
--image-url "https://example.com/photo.jpg" \
--prompt "Convert to anime style" \
--operation style
# Text to speech
bash skills/claude.ai/fal-audio/scripts/text-to-speech.sh \
--text "Hello world"
# Search for models
bash skills/claude.ai/fal-generate/scripts/search-models.sh \
--query "text to video"
Each skill follows the community skills standard:
claude.ai/
└── skill-name/
├── SKILL.md # Metadata + documentation (YAML frontmatter)
└── scripts/
└── script.sh # Executable bash script
The SKILL.md contains a YAML frontmatter with name, description, and metadata, followed by usage documentation that the agent reads to understand how to use the skill.
MIT