Media Generation Overview
Elftia integrates multiple AI media generation capabilities, covering image, video, and music creation as well as speech recognition (ASR) and speech synthesis (TTS), so you can generate multimedia content with AI directly within your chat and creation workflows.
Use Cases
- Generate illustrations, concept designs, or artwork with AI during a chat
- Generate short videos from text descriptions
- Create AI music from text prompts and optional lyrics
- Automatically transcribe voice messages to text
- Convert AI replies or custom text to speech for playback
Supported Media Types
| Media type | Description | Supported providers |
|---|---|---|
| Image generation | Generate images from text prompts | Gemini, OpenAI (DALL-E), Vertex AI, custom providers |
| Video generation | Generate video from text or images | Kling AI, OpenAI (Sora), Replicate, VolcEngine, Wan |
| Music generation | Create music from text prompts | ACEMusic |
| Speech recognition (ASR) | Transcribe audio files to text | OpenAI Whisper, ElevenLabs Scribe, Fish Audio |
| Speech synthesis (TTS) | Convert text to speech | ElevenLabs, Fish Audio |
General Workflow
All media generation features follow a similar workflow:
Configure provider → Select model → Enter parameters → Generate → View / save result
Step 1: Configure a Provider
- Open Settings > Media Providers
- Find the provider for the media type you want to use
- Enter the API key and any required configuration
- Enable the provider
Step 2: Generate Content
Depending on the media type, enter the parameters in the corresponding UI and generate the content. See the detailed documentation for each media type for specific steps.
Step 3: View and Save
Generated media content can be previewed directly and saved to the local file system.
Global Settings
All media generation features share the following global settings:
| Setting | Description | Default |
|---|---|---|
| Concurrency | Maximum number of simultaneous generation tasks | 3 |
| Retry count | Number of automatic retries when generation fails | 2 |
Provider Management
Built-in Providers
Elftia includes definitions for the major media generation providers. You only need to enter an API key to get started.
Custom Providers
For providers not in the built-in list, you can add a custom provider manually:
- Click Add Custom Provider in the media provider settings
- Fill in the provider name, API endpoint, and authentication information
- Configure the model list and default parameters
- Save and enable
Detailed Documentation
- Image Generation — Complete guide for generating images with AI
- Video Generation — Instructions for generating video with AI
- Music Generation — Parameter configuration for creating music with AI
- Speech Services — Speech recognition and speech synthesis features
Frequently Asked Questions
| Question | Solution |
|---|---|
| Provider shows as not configured | Check that the correct API key has been entered |
| Generation request times out | Some generation tasks (especially video) take a long time — please wait |
| Poor generation quality | Try refining the prompt description; add more detail and style guidance |
| API quota exhausted | Check your usage quota with the provider; consider upgrading your plan or switching providers |
Related Links
- LLM Provider Configuration — How to configure LLM text models (separate from media providers)
- Chat — Using media generation features within a conversation