Fal ai media generation
Skill findscripter/everything-skills/06-creative/fal-ai-media-generation
当需要用 AI 生成图像、视频或语音/音乐/音效时使用;经 fal.ai MCP 的 search/find/generate/result/upload 等工具运行 Nano Banana、Seedance、Kling、Veo 3、CSM-1B、ThinkSound 等模型,产出图片/视频/音频文件或 URL;不适用于未配置 fal.ai MCP(FAL_KEY)或需精细本地剪辑调色的场景,后者回退 ffmpeg/VideoDB;触发词:生成图片、文生图、做缩略图、文生视频、图生视频、文字转语音、视频配音From its SKILL.md
npx -y skills add findscripter/everything-skills --skill fal-ai-media-generationAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its file declares
Copied from the file, not written here
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
7.1 KB, ~2.0k tokens by cl100k_base, as published. Nobody here has run it
fal.ai 多模态媒体生成(图像/视频/音频)
经 fal.ai MCP 服务器统一调度图像、视频、音频模型:文生图(Nano Banana)、文/图生视频(Seedance、Kling、Veo 3)、文字转语音(CSM-1B)、视频转音频(ThinkSound)。一切生成在 fal.ai 服务端完成,本地只下发参数、取回结果 URL。
何时使用
适用:
- 从文本提示生成图像;做缩略图、海报、产品图;对已有图做重绘/风格迁移/局部编辑。
- 从文本或图像生成视频;为视频生成配音/音效;文字转语音、生成音乐/音效。
- 用户说「生成图片」「做个缩略图」「文生视频」「图生视频」「文字转语音」「给这段视频配音」等。
不该用(负边界):
- 未配置 fal.ai MCP(无
FAL_KEY):所有generate调用都不可用,先让用户自行配置 key,切勿由你读写 key。 - 需要精细本地剪辑(转场、变速、裁切调色、多轨混音、时间轴拼接):fal.ai 只做「生成」,这类编辑回退 ffmpeg /
videodb-perception-editing。 - 专业级人声克隆/配音:用 ElevenLabs 直连 API(见示例),不走 MCP。
步骤
- 确认 MCP 已配置:
~/.claude.json加入 fal-ai server(见下)。无 key 则停下来请用户提供。 - 选模型:迭代用低价快模型(Nano Banana 2 / Seedance),定稿换高保真(Nano Banana Pro)。不确定就先
search/find/models发现。 - 贵活先估价:视频等高成本任务先
estimate_cost。 - 需以图/视频为输入时先
upload拿 URL,再把 URL 填进image_url/video_url。 - 运行
generate;异步任务用status/result轮询,必要时cancel。 - 取回结果 URL / 文件,交付或转入后续编辑流程。
指令
MCP 配置(~/.claude.json,key 申请见 https://fal.ai ):
"fal-ai": {
"command": "npx",
"args": ["-y", "fal-ai-mcp-server"],
"env": { "FAL_KEY": "YOUR_FAL_KEY_HERE" }
}
MCP 工具:search(按关键词找模型)/ find(看模型参数)/ generate(运行)/ result、status(查异步状态)/ cancel(取消)/ estimate_cost(估价)/ models(热门模型)/ upload(上传输入文件)。
图像通用参数:prompt(必填) / image_size(square,portrait_4_3,landscape_16_9,portrait_16_9,landscape_4_3) / num_images(1-4) / seed(整数,复现用) / guidance_scale(1-20,越高越贴提示)。
视频通用参数:prompt(必填) / duration("5s","10s") / aspect_ratio("16:9","9:16","1:1") / seed / image_url(图生视频源图)。
选型速查:
- 图像快出/编辑 →
fal-ai/nano-banana-2;高保真/排版/写实/产品图 →fal-ai/nano-banana-pro。 - 视频高动态质量 →
fal-ai/seedance-1-0-pro;带原生音频 →fal-ai/kling-video/v3/pro;带生成音效+高画质 →fal-ai/veo-3。 - 语音 →
fal-ai/csm-1b;视频配音 →fal-ai/thinksound。
示例
文生图(快速迭代):
generate(model_name: "fal-ai/nano-banana-2", input: {
"prompt": "a futuristic cityscape at sunset, cyberpunk style",
"image_size": "landscape_16_9", "num_images": 1, "seed": 42
})
高保真产品图:
generate(model_name: "fal-ai/nano-banana-pro", input: {
"prompt": "professional product photo of wireless headphones on marble surface, studio lighting",
"image_size": "square", "num_images": 1, "guidance_scale": 7.5
})
图像编辑(先 upload 再带 image_url):
upload(file_path: "/path/to/image.png")
generate(model_name: "fal-ai/nano-banana-2", input: {
"prompt": "same scene but in watercolor style",
"image_url": "<uploaded_url>", "image_size": "landscape_16_9"
})
文生视频 / 图生视频:
generate(model_name: "fal-ai/seedance-1-0-pro", input: {
"prompt": "a drone flyover of a mountain lake at golden hour, cinematic",
"duration": "5s", "aspect_ratio": "16:9", "seed": 42
})
# 图生视频:再补 "image_url": "<uploaded_image_url>"
文字转语音 / 视频转音频:
generate(model_name: "fal-ai/csm-1b", input: {
"text": "Hello, welcome to the demo.", "speaker_id": 0
})
generate(model_name: "fal-ai/thinksound", input: {
"video_url": "<video_url>", "prompt": "ambient forest sounds with birds chirping"
})
专业人声(ElevenLabs 直连,不走 MCP):
import os, requests
resp = requests.post(
"https://api.elevenlabs.io/v1/text-to-speech/<voice_id>",
headers={"xi-api-key": os.environ["ELEVENLABS_API_KEY"], "Content-Type": "application/json"},
json={"text": "Your text here", "model_id": "eleven_turbo_v2_5",
"voice_settings": {"stability": 0.5, "similarity_boost": 0.75}})
open("output.mp3", "wb").write(resp.content)
发现模型:search(query: "text to video") / find(model_name: "fal-ai/seedance-1-0-pro") / models()。
注意事项
- 无
FAL_KEY一切免谈:先确认 MCP 配好,key 由用户提供,你不要读写它。 - 迭代提示词用
seed复现;先低价模型调通提示,再换 Pro 出最终图。 - 视频提示词描述要聚焦「运动 + 场景」,简洁优先;图生视频比纯文生视频更可控。
- 跑昂贵视频前务必
estimate_cost。 - 异步任务别死等,用
status/result轮询,跑偏就cancel。 - 精细剪辑/调色/混音不是 fal.ai 的活,回退本地 ffmpeg 或 VideoDB。
互见
- related:
videodb-perception-editing—— 视频感知索引与服务端时间轴编辑,承接 fal.ai 生成后的剪辑/字幕/检索。 - related:
demo-video-generator—— 从零渲染演示视频,与本技能的纯 AI 生成互补。 - combines_with:
videodb-perception-editing—— fal.ai 出素材 + VideoDB 编排时间轴,串成完整媒体流水线。
采编自 affaan-m/everything-claude-code(MIT)。
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.
Gives 0 of the 12 instructions most media documents skills give in ~2.0k tokens
Counted across 157 of the 158 authors here whose files we hold, read 2026-08-07
- Provide posting time recommendationsin 7 of 157, across 5 files
- Track metrics over time to identify trendsin 6 of 157, across 2 files
- Read marketing context file before startingin 6 of 157, across 5 files
- Choose platforms based on audience presencein 6 of 157, across 4 files
- Adapt tone for each platformin 6 of 157, across 4 files
- Ensure data completeness before analysisin 5 of 157, across 1 file
- Compare metrics within same time periodsin 5 of 157, across 1 file
- Account for platform-specific benchmarksin 5 of 157, across 1 file
- Separate organic and paid metricsin 5 of 157, across 1 file
- Include context when interpreting resultsin 5 of 157, across 1 file
- Keep tweets under 280 charactersin 5 of 157, across 3 files
- Download top-K results with an attribution sidecarin 5 of 157, across 2 files
Said here and by no other author read
- stop if fal key is missing
- request user to configure key
- do not read or write key
- use high fidelity models for final output
- upload input media before generation
- poll status for asynchronous tasks
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.