用 n8n 自动编辑并发布 TikTok 视频：Gemini Omni Flash + OpenAI

输入视频 URL 与编辑指令 → 经 Fal Gemini Omni Flash 渲染 → OpenAI 生成标题 → 上传至 TikTok（Upload-Post）

给谁

需要批量处理视频并自动发布到 TikTok 的内容创作者或运营团队

你会得到

通过 webhook 触发完整流水线：AI 编辑视频、自动生成 SEO 标题、自动上传发布

你需要

n8n

步骤

1. 触发工作流并提交视频编辑任务

   在 n8n 中配置 Webhook 节点，设置触发路径。确保外部调用以 POST 发送 JSON，包含 `video_url`（源视频地址）与 `prompt`（编辑指令，如"Make this video anime"）。Webhook 接收后，数据流入 Create Video 节点。该节点向 `https://queue.fal.run/google/gemini-omni-flash/edit` 发送 POST 请求，请求体为 JSON，包含上述 prompt 与 video_url。需提前在 n8n Credentials 中添加 Fal.run API 的 httpHeaderAuth 凭证并在节点中选中。

   运行提示词 · 来源原文。Create Video 节点的 JSON Body 字段；需开启 Send Body、Specify Body 设为 JSON

   复制提示词

   ={
        "prompt": "{{ $json.prompt }}",
        "video_url": "{{ $json.video_url }}"
      }

2. 轮询渲染状态直到完成

   Create Video 返回后，工作流进入 Wait 60 sec. 节点等待60秒，随后 Get status 节点向 Fal.run 查询任务状态（GET `https://queue.fal.run/google/gemini-omni-flash/requests/{{request_id}}/status`）。Completed? 节点检查返回状态是否为 `COMPLETED`：若是，进入下一步；若否，回退到 Wait 60 sec. 再次等待。循环直至渲染完成。注意：request_id 来自 Create Video 的返回。

3. 生成标题、下载视频并上传至 TikTok

   渲染完成后，Get Url Video 节点获取最终视频详情（GET `https://queue.fal.run/google/gemini-omni-flash/requests/{{request_id}}`），提取 `video.url`。随后 Generate title 节点调用 OpenAI（模型 gpt-5），以系统提示词要求生成 SEO 友好标题（≤60字符，与输入同语言），用户消息为原始 prompt。接着 Get File Video 节点从该 URL 下载视频文件。最后 HTTP Request 节点向 `https://api.upload-post.com/api/upload` 发送 POST（multipart/form-data），包含：title（来自 Generate title 输出）、user（替换为实际 Upload-Post 用户名）、platform[]（tiktok）、video（二进制数据，字段名 data）。需配置 Upload-Post 的 httpHeaderAuth 凭证。

   运行提示词 · 来源原文。Generate title 节点的 System Message 字段

   复制提示词

   You are a Video SEO expert specialized in creating engaging and optimized titles.
   
   Your task is to generate an effective title for a YouTube video based on the user's video description.
   
   GUIDELINES:
   - Maximum 60 characters to avoid truncation
   - Use relevant keywords for SEO
   - Make the title catchy and clickable
   - Avoid excessive or misleading clickbait
   - Consider the target audience of the content
   - Use numbers, questions, or power words when appropriate
   - IMPORTANT: Generate the title in the same language as the input description
   
   OUTPUT FORMAT:
   Provide only the title, without additional explanations.
   
   EXAMPLE:
   Input: "Tutorial video on how to cook perfect pasta carbonara"
   Output: "PERFECT Carbonara in 10 Minutes - Chef's Secrets"

   运行提示词 · 来源原文。Generate title 节点的 User Message 字段（Content），需将 Role 设为 user

   复制提示词

   =Input: {{ $('Webhook').item.json.prompt }}

4. 返回上传结果给调用方

   HTTP Request（上传）节点完成后，Respond to Webhook 节点将 Upload-Post API 的响应原样返回给最初触发 webhook 的调用方。至此流水线结束。

# Edit Video with Gemini Omni Flash and upload to TikTok

## How It Works

1. A webhook receives a video editing request and starts the workflow.
2. The workflow submits the request to Gemini Omni Flash to edit the video and waits for the rendering process to complete.
3. It continuously checks the job status until the video has been successfully rendered.
4. Once the job is complete, the workflow retrieves the generated video URL, creates an SEO-friendly title using OpenAI, and downloads the final video.
5. Finally, the workflow uploads the processed video to Upload-Post and returns the upload response to the original caller.

## Setup Steps

- [ ] Configure the **Webhook** trigger path and ensure the incoming payload includes both the `video_url` and `prompt` fields.
- [ ] Configure your **Fal API** credentials for the HTTP Request nodes used to submit the editing job, check the rendering status, and retrieve the final video information.
- [ ] Verify that the **Post Edit Job to Gemini** node contains the correct editing prompt and the source video URL.
- [ ] Configure your **OpenAI** credentials for the **Generate Video Title** node.
- [ ] Ensure the **Download Video File** node can access the signed URL returned by the Fal API.
- [ ] Configure your **Upload-Post API** credentials in the **Upload Video to Upload-Post** node.

Create an [Upload-Post account for FREE](https://www.upload-post.com/?linkId=lp_144414&sourceId=n3witalia&tenantId=upload-post-app) and obtain your API credentials visit:

## Customization

You can easily customize this workflow by:

- Changing the polling interval in the **Wait 60 Seconds** node.
- Modifying the prompt used by the **Generate Video Title** node.
- Updating the default upload parameters (user, platforms, metadata, etc.) sent to Upload-Post.
- Replacing the final upload destination with another storage service or API if required.