OpenCreator Docs

Install the app, connect an AI service, and complete your first creation. Find settings, provider values, and fixes here.

Download OpenCreator

Get started

Download the installer for your computer, then configure the AI service you plan to use.

  1. Install the appDownload the installer for your Mac or Windows computer.
  2. Find the settingsOpen Settings → AI Services and use the screenshots to locate the fields.
  3. Enter provider detailsChoose a service in the provider reference and enter its Base URL, API Key, and model. Add a proxy URL if your network needs one.
  4. Run a real testFollow Verify your setup. A text reply or an image you can preview and download confirms that service works. If the request fails, see Troubleshooting.
Mac

Choose the .dmg installer for your Mac's chip.

.dmg
Windows

Download the Windows .exe installer.

.exe
After installation

Open OpenCreator and go to Settings → AI Services. Next: find the fields in the screenshots.

Find settings in OpenCreator

Open Settings in the left sidebar, then select AI Services. In Models, enter the Base URL, API Key, and Model; the Network area below has the Proxy URL field. Use the tabs across the top for image, voice, transcription, and video services.

OpenCreator left sidebar with Settings selected and AI Services in the settings menu
Navigation: Settings → AI Services
OpenCreator AI Services, showing the Models tab and Base URL, API Key, and Model fields
AI Services → Models. Open the image to see it at full size. The api.example.com address and model shown are illustrative; use the provider reference below for actual values.

Next: find the values for your provider, or set a network proxy if needed.

Set a network proxy URL

If your network cannot reach a required AI service or media address directly, enter an HTTP or HTTPS proxy you already have. The proxy routes model and media requests. It is separate from the provider's Base URL and does not replace its API Key.

Network section below Models in OpenCreator AI Services, with the Proxy URL field and Save settings button
Settings → AI Services → Models → Network. The proxy.example.com address in the screenshot is illustrative; enter your own proxy URL.
  1. Make sure the proxy is runningGet its HTTP/HTTPS address and port from your proxy tool or network administrator. For a proxy running on this computer, an address may look like http://127.0.0.1:7890; use the port your tool actually provides.
  2. Enter it in OpenCreatorOpen Settings → AI Services → Models, find Network → Proxy URL, enter the full http:// or https:// URL, and select Save settings. This field is optional; socks5:// URLs are not accepted.
  3. Verify the connectionRun a small test. If it still times out, check that the proxy is running and its address and port are correct, then check the provider Base URL, model, and API Key.

yt-dlp video extractor

OpenCreator uses yt-dlp to analyze and download public videos from supported platforms such as YouTube and Bilibili. If you use a platform video URL, open Settings → Third-party Components and check the yt-dlp nightly status. Importing a local video file does not require this step.

OpenCreator Third-party Components showing the yt-dlp nightly version, Check for updates, and Update to buttons
Settings → Third-party Components → yt-dlp nightly. Version numbers and update status depend on your installation.
  1. Check the current versionFind yt-dlp nightly in Third-party Components. A current version number means the local Runtime found the component; you do not need to install it separately in a terminal.
  2. Check and update when neededSelect Check for updates. If a newer version appears, select Update to…; checking alone does not install it. If a platform changes and a URL stops working, update yt-dlp and retry.
  3. Test a video URLOpen Video Downloader from the Dashboard and submit an accessible public video URL. If OpenCreator cannot reach the platform, check the URL and your network proxy first. If the extractor may be outdated, check for a yt-dlp update.

If the update download fails, check your network and proxy URL. A failed verification or save leaves the current working version available.

Understand the configuration fields

OpenCreator connects to external AI services. These four fields tell it which service to call and how.

01
Provider

Which company provides the AI service?

02
Base URL

Where should OpenCreator send requests?

03
API Key

Which credential authorizes the request?

04
Model

Which model should the service run?

Example: a DeepSeek text model
Provider
DeepSeek
The AI service you chose.
Base URL
https://api.deepseek.com
The service's root address.
API Key
Your DeepSeek API Key
Created in the DeepSeek platform.
Model
deepseek-v4-pro
The specific model ID to use.

Provider reference

Find a service by task, then use its detailed entry below. Model IDs are examples; use an ID available to your account if the provider's catalog has changed.

TaskProviderBase URLExample model IDGet an API Key
TextOpenAIhttps://api.openai.com/v1gpt-4o-miniOpenAI Platform ↗
TextDeepSeekhttps://api.deepseek.comdeepseek-v4-proDeepSeek Platform ↗
ImageOpenAIhttps://api.openai.com/v1gpt-image-1OpenAI Platform ↗
ImageGeminihttps://generativelanguage.googleapis.com/v1betagemini-2.5-flash-imageGoogle AI Studio ↗
VideoSeedance / Volcano Arkhttps://ark.cn-beijing.volces.com/api/v3doubao-seedance-2-5-260628Volcano Ark Console ↗

OpenAI

For text and image generation. Both configurations can use the same OpenAI API Key.

Get an API Key ↗
In OpenCreator
Settings → AI Services. For text, select OpenAI under the custom model service; for image generation, select OpenAI under Images.
Base URL
https://api.openai.com/v1
Model IDs
Text: gpt-4o-mini   Image: gpt-image-1

Choose the model for the task you configure. Official model catalog ↗

Where to get the key
Sign in to OpenAI Platform → API keys → Create new secret key. Paste it into the API Key field for the service.

DeepSeek

A text model service you can use without an OpenAI text model.

Get an API Key ↗
In OpenCreator
Settings → AI Services → custom model service → DeepSeek.
Base URL
https://api.deepseek.com
Model ID
deepseek-v4-pro

Use an ID available in your DeepSeek account. Official model reference ↗

Where to get the key
Sign in to DeepSeek Platform → API Keys → Create API Key. Enter it in OpenCreator's API Key field.

Gemini

An example service for image generation.

Get an API Key ↗
In OpenCreator
Settings → AI Services → Images → Gemini.
Base URL
https://generativelanguage.googleapis.com/v1beta
Model ID
gemini-2.5-flash-image

Official model catalog ↗

Where to get the key
Open Google AI Studio → Get API key → Create API key. Enter it in the Gemini API Key field.

Volcano Ark / Seedance

An example service for video generation. Use a Volcano Ark API Key.

Open the console ↗
In OpenCreator
Settings → AI Services → Video → Seedance.
Base URL
https://ark.cn-beijing.volces.com/api/v3
Model ID
doubao-seedance-2-5-260628

Use a model ID enabled and available in your Volcano Ark console.

Where to get the key
Sign in to Volcano Ark Console → API Key Management → Create API Key. Make sure your account has access to the model.

After saving: run a small test

Two configuration examples

Choose either combination, or configure only the service categories you need.

Example one: with OpenAI

Use OpenAI for text and images, and Seedance for video.

CategoryProviderBase URLModelAPI Key
TextCustom model service → OpenAIhttps://api.openai.com/v1gpt-4o-miniOpenAI API Key
ImageOpenAIhttps://api.openai.com/v1gpt-image-1The same OpenAI API Key
VideoSeedancehttps://ark.cn-beijing.volces.com/api/v3doubao-seedance-2-5-260628Volcano Ark API Key

Example two: without an OpenAI text model

Use other AI services for text, images, and video.

CategoryProviderBase URLModelAPI Key
TextCustom model service → DeepSeekhttps://api.deepseek.comdeepseek-v4-proDeepSeek API Key
ImageGeminihttps://generativelanguage.googleapis.com/v1betagemini-2.5-flash-imageGemini API Key
VideoSeedancehttps://ark.cn-beijing.volces.com/api/v3doubao-seedance-2-5-260628Volcano Ark API Key

Verify your setup

Saving settings only stores the values. Make a real request in the matching tool to verify the API Key, model, and address.

Configured serviceSmall testSuccess looks like
Text modelSelect the configured model in a conversation and send a short request.You receive a normal reply without a provider error.
Image serviceOpen Image Generation from the Dashboard and generate one simple image.The result appears and can be previewed or downloaded.

Test only the services you need. Each provider key and model must be verified separately. If a request fails, see Troubleshooting.

Test passed? Start creating

Before you save

Use the Base URL as shown

For example, enter https://api.deepseek.com exactly as listed in the provider reference.

Correct:
https://api.deepseek.com

Do not enter:
https://api.deepseek.com/chat/completions

OpenCreator calls the specific API endpoint itself. Enter only the service's root address.

Do not mix API Keys from different providers

Each key belongs to the company that issued it.

An OpenAI API Key cannot call DeepSeek.
A DeepSeek API Key cannot call Gemini.
A Gemini API Key cannot call Seedance.

Match each Base URL with an API Key from the same provider.

Creator guides

Open a tool from the Dashboard. These two workflows use a configured image service or the model and transcription services needed for video translation.

Cropped OpenCreator Dashboard tool list with Video Translation and Image Generation
Cropped Dashboard tool list: Video Translation and Image Generation.

Generate your first image

Configure an image service in AI Services → Images first. For OpenAI, use the OpenAI reference.

  1. Open the Dashboard and choose Image Generation.
  2. Enter a prompt. Optionally add a reference image and choose the aspect ratio and output count.
  3. Preview and download the result. If generation fails, check the image provider's key, model, and quota.

Translate a video

Prepare a video and configure a text model and transcription service. Configure a voice service if you also need dubbing. For supported online video URLs, check the yt-dlp status first.

  1. Open the Dashboard, choose Video Translation, and import a local video or accessible video URL.
  2. Choose the source and target languages, review the transcript and subtitles, and adjust dubbing or subtitle styles if needed.
  3. Preview the result, then export subtitles, audio, or video. If the task stops at transcription, check that service first.

Request failed? Find the symptom and fix

Troubleshooting

Start with the symptom shown by OpenCreator or the provider, then retry a small task. Never share an API Key in screenshots or support requests.

Authentication failed or 401 / 403

Use a key from the provider behind the current Base URL. Check that your account can call the selected model, then re-enter and save the key.

Model not found or 404

Check the model ID against your account's available models. Enter only the service root as the Base URL, without an endpoint such as /chat/completions.

Quota exceeded, rate limited, or 429

Check balance, quota, and rate limits in the provider console, then retry later. Make sure the chosen model is enabled for your account.

Connection timed out

Retry with a short request. If your network needs a proxy, follow the network proxy setup and check that it is running with the right address and port. Then check access to the provider address.

Online video URL cannot be analyzed or downloaded

Confirm the URL is accessible and the platform is supported. Check your proxy URL if the network requires one. If the platform has changed its rules, check and update yt-dlp in Third-party Components, then retry.

Settings saved, but no result appears

Check that the tool uses the service you configured, then follow the small tests. Text, image, transcription, and voice services are configured separately.

The setup in one glance
Provider
Whose AI service?
Base URL
Where to reach it?
API Key
Which credential?
Model
Which specific AI?