| 1 | # Configuration |
| 2 | |
| 3 | After installation, you need to configure services to use Pixelle-Video. |
| 4 | |
| 5 | --- |
| 6 | |
| 7 | ## LLM Configuration |
| 8 | |
| 9 | LLM (Large Language Model) is used to generate video scripts. |
| 10 | |
| 11 | ### Quick Preset Selection |
| 12 | |
| 13 | 1. Select a preset model from the dropdown: |
| 14 | - Qianwen (recommended, great value) |
| 15 | - GPT-4o |
| 16 | - DeepSeek |
| 17 | - Ollama (local, completely free) |
| 18 | |
| 19 | 2. The system will auto-fill `base_url` and `model` |
| 20 | |
| 21 | 3. Click「🔑 Get API Key」to register and obtain credentials |
| 22 | |
| 23 | 4. Enter your API Key |
| 24 | |
| 25 | --- |
| 26 | |
| 27 | ## Image/Video Generation Configuration |
| 28 | |
| 29 | Two options available: |
| 30 | |
| 31 | ### Local Deployment |
| 32 | |
| 33 | Using local ComfyUI service: |
| 34 | |
| 35 | 1. Install and start ComfyUI |
| 36 | 2. Enter ComfyUI URL (default `http://127.0.0.1:8188`) |
| 37 | 3. Click "Test Connection" to verify |
| 38 | 4. (Optional) Enter ComfyUI API Key (get from [Comfy Platform](https://platform.comfy.org/profile/api-keys)) |
| 39 | |
| 40 | ### Cloud Deployment (Recommended) |
| 41 | |
| 42 | Using RunningHub cloud service, no local GPU required: |
| 43 | |
| 44 | 1. Register for a RunningHub account |
| 45 | 2. Obtain API Key |
| 46 | 3. Enter API Key in configuration |
| 47 | 4. Configure advanced options (optional): |
| 48 | - **Concurrent Limit**: Set number of simultaneous tasks (1-10, default 1 for regular members) |
| 49 | - **Instance Type**: Choose 24GB or 48GB VRAM machine (48GB for large video generation) |
| 50 | |
| 51 | --- |
| 52 | |
| 53 | ## Save Configuration |
| 54 | |
| 55 | After filling in all required configuration, click the "Save Configuration" button. |
| 56 | |
| 57 | Configuration will be saved to `config.yaml` file. |
| 58 | |
| 59 | --- |
| 60 | |
| 61 | ## Next Steps |
| 62 | |
| 63 | - [Quick Start](quick-start.md) - Create your first video |
| 64 | |
| 65 |