OpenAI’s Sora can generate videos from text prompts. Seriously impressive stuff. But the API’s a bit clunky to work with directly.

That’s why we built the Sora MCP bundle. 6 tools that let you create, remix, and manage AI-generated videos without touching code. Just tell the AI what you want and it handles the rest.

This video was generated using the Sora MCP bundle—we used our own tools to create it.

What You Can Actually Do​

Create videos from scratch. Describe what you want: “A cat playing piano in a jazz club” or “Ocean waves at sunset with seagulls flying overhead.” Sora generates a video that matches your prompt. Takes 2-5 minutes depending on length.

Remix existing videos. Got a video you like but want to change something? “Make it more colorful” or “Add rain to this scene.” Sora remixes your video based on the new prompt.

Check status while videos generate. Video generation isn’t instant—it takes a few minutes. Check the status to see if it’s queued, processing, or ready to download.

Download completed videos. Once your video’s done, download it as an MP4. We automatically store it in S3 and give you a download link that works for an hour. No API keys needed.

List all your videos. See everything you’ve generated, when it was created, and its current status.

Delete videos you don’t need. Clean up old videos or ones that didn’t turn out right.

The One-Hour Download Window​

Here’s something important: OpenAI only lets you download videos for one hour after they’re completed. After that, the download link expires and the video’s gone forever.

This isn’t a bug—it’s OpenAI’s policy. They confirmed it in their API responses and error messages. Videos expire exactly one hour after completion.

Our tools handle this automatically. When you download a video, we grab it from OpenAI immediately, upload it to secure S3 storage, and give you a presigned download link that works without exposing your API key. You’ve got permanent access as long as you download within that first hour.

After the hour? You’ll need to regenerate the video. That’s just how Sora works.

How It Actually Works​

Say you want a video of a robot dancing. You ask: “Create a video of a friendly robot dancing in a neon-lit room.”

Behind the scenes, the AI uses sora_create_video with your prompt. It sets reasonable defaults for video length and resolution unless you specify something different. Sora starts generating the video and returns a video ID.

Then it polls the status using sora_get_video_status until the video’s ready. This takes 2-5 minutes typically. You can wait or come back later—the AI tracks the video ID either way.

Once it’s done, the AI automatically downloads the video using sora_download_video. We pull the video from OpenAI’s API, upload it to S3 under your user account, and return a presigned download URL that expires in one hour. Click the link and you’ve got an MP4 file.

Want to remix it? “Make it more colorful and vibrant.” The AI uses sora_remix_video with your original video ID and the new prompt. Generates a new video with your modifications.

Video Options You Can Control​

Length: 4, 8, 12, 16, or 20 seconds. Longer videos take longer to generate but give you more content.

Resolution: Choose from 720x1280 (vertical mobile), 1280x720 (horizontal desktop), 1080x1920 (vertical HD), or 1920x1080 (horizontal HD). Pick based on where you’re using the video.

Model: Currently just sora-2, but OpenAI will likely add more models over time.

You don’t need to specify any of this if you don’t want to. The AI picks sensible defaults based on your request.

Real Use Cases​

Social media content: Generate short videos for Instagram, TikTok, or Twitter without filming anything. “A coffee cup steaming on a desk with morning sunlight.”

Concept visualization: Show clients what an idea might look like before investing in production. “A modern office lobby with plants and natural light.”

Creative experimentation: Try ideas that would be expensive or impossible to film. “A city street transforming from day to night in fast motion.”

Product demos: Generate videos showing how something works. “A smartphone displaying a weather app with animated clouds.”

The quality’s good enough for real use. Not perfect—AI video generation still has quirks—but impressive enough that people won’t immediately know it’s AI-generated.

Getting Started​

Enable the Sora MCP bundle in your MCPBundles account. You’ll need an OpenAI API key with Sora access. Get one from platform.openai.com/api-keys.

Add your API key once during setup. Then start creating videos. The AI automatically picks the right tools based on what you ask.

“Create a video of a sunset over mountains” → Creates the video “Check if my video is ready” → Checks status “Download the video” → Gets you the MP4 file

All the technical stuff—API calls, status polling, S3 uploads, download links—happens automatically.

What About Costs?​

Sora isn’t cheap. Video generation costs vary based on length and resolution. Expect to pay $0.50-$2.00 per video depending on what you generate.

Our MCP tools don’t add any cost on top of OpenAI’s pricing. You pay OpenAI directly for video generation.

The automated downloads to S3 use your own AWS credentials, so storage costs are minimal (pennies per gigabyte per month).

Limitations to Know About​

One-hour download window: Download your videos within an hour or they’re gone. This is OpenAI’s policy, not ours.

Generation time: Videos take 2-5 minutes to generate. You can’t speed this up—it’s how long Sora needs to process the video.

Quality varies: AI video generation is impressive but not perfect. Some prompts work better than others. Expect some trial and error.

API access required: You need OpenAI API access with Sora enabled. Not everyone has this yet—it’s still rolling out.

Start Creating Videos​

Enable the Sora MCP bundle and connect your OpenAI API key. Then describe what you want. The AI handles the rest.

No video editing skills needed. No API documentation to read. Just prompts and results.