How to Use Dreamscape

Ritual Dreamscape is a suite of AI applications utilizing decentralized TEE (Trusted Execution Environment) co-processors on the Ritual Chain. Here is how to use the specific tools within this application.

1Funds: Locking & Withdrawing

To prevent griefing and spamming of the network, the RitualWallet requires you to lock funds before executing asynchronous AI generations.

  • Deposit & Lock: You specify an amount of RITUAL tokens and a duration (in blocks). During this time, your funds act as collateral. If a task fails or you behave maliciously, slashing could occur (though primarily applicable to executors).
  • Withdraw: Once your lock duration expires (i.e., the current block number exceeds your lock block), you can withdraw your RITUAL tokens back to your standard wallet balance.

Before running any AI generation, ensure you have sufficient balance and your funds are deposited in the Funds page.

2LLM Prompt Optimization

Using the Ritual LLM Precompile, we can transform simple, raw ideas into highly optimized prompts specifically formatted for Image, Audio, and Video models.

  • Navigate to the Prompt Optimization tool.
  • Enter a short description (e.g., "Cyberpunk city in the rain").
  • The DApp securely sends this to the LLM (GLM-4.7) running in a TEE.
  • The model acts as an expert FLUX prompt engineer and returns a highly detailed JSON image prompt optimized for generation.

3Image Generation

Transform your text prompts into stunning, high-definition AI artworks.

  • Head to the Image Tool.
  • Paste the visual prompt generated in step 2 (or write your own).
  • Click Generate. The AI will securely process your request and return a gorgeous image.
  • Once it loads, you can download the artwork directly to your device!

4Audio Generation

Similar to Image Generation, Ritual provides an Audio Precompile. We utilize the LiquidAI model to power natural, conversational audio chats.

  • Visit the Audio Tool and type a prompt or question into the text box.
  • Because the LiquidAI model is a Conversational AI, you can ask it questions and it will answer you naturally!
  • Click Generate. The AI will securely synthesize its spoken response and deliver an audio player right on your screen.
  • You can play the audio directly in your browser or download it to save it!

5Video Generation

Generate short cinematic video clips using the powerful Wan2.2 model right from the blockchain.

  • Visit the Video Tool and type your visual prompt into the text box.
  • Click Generate. Because video takes massive computational power, the TEE executor will take longer to compile it (~120-240 seconds).
  • The smart contract handles the extended wait time seamlessly, and your video will appear on screen once complete.
  • You can play the video directly in your browser or download the MP4 file!