Earning module
Opt in and your GPU joins the dispatch pool when idle. AI inference workloads are routed to your machine; you’re paid in GUSD per completed job.- Earn while idle. Workloads use the full GPU and VRAM; launching a game pushes the module out automatically. Just leave it on whenever you’re not gaming.
- Pause anytime. No minimum commitment — toggle off in the sidebar.
Home module
The dashboard that runs in the background and ties your account to network rewards.- GHXP accrual — every minute the app is running, you accrue GHXP. Pending points credit at the top of each hour. Hold GHX in MetaMask or your in-app wallet to multiply the rate.
- Seasons — every two months, lock your accumulated GHXP to claim a proportional share of the season’s GHX pool.
- Account status — current GHXP, level, season standing, and connected wallets, all on one screen.
Image
| Module | What it does | Models |
|---|---|---|
| Text-to-Image | Generate images from prompts. Results stay locally; the app keeps a history per device. | Flux.1 dev, FLUX.1 Schnell 12B, FLUX.2 Klein 4B, Z-Image-Turbo, Sana Sprint 0.6B / 1.6B |
| Image Edit | Edit an existing image with a text instruction — change a subject, swap a background, restyle, inpaint. | Qwen-Image-Edit Plus |
| Image Upscale | Increase resolution of an existing image without re-prompting. | RealESRGAN x2 / x4 |
| Background Removal | Cut subjects out — clean alpha output for compositing or transparent assets. | Ben2 |
Video
| Module | What it does | Models |
|---|---|---|
| Image / Text / Audio → Video | Turn a still image, a text prompt, or an audio track into a short video clip. | LTX-2.3 22B Distilled, LTX-2 19B Distilled, LTX-Video 0.9.8 13B |
| Video Upscale | Increase resolution of an existing video clip. | FlashVSR Tiny, RealESRGAN Video x2 / x4 |
| Video Replace / Animate | Animate or replace a subject in an existing video — swap a character, drive motion from a reference, restyle the action. | Wan2.2-Animate 14B |
Voice & Audio
| Module | What it does | Models |
|---|---|---|
| Text-to-Speech | Generate speech from text — design a custom voice, clone an existing one, or use a default. | Qwen3 TTS (VoiceDesign / VoiceClone / CustomVoice), Chatterbox, Kokoro |
| Voice & Audio Transcription | Transcribe speech to text from audio or video — files or live streams. | Whisper Large V3 |
| Text-to-Music | Generate music tracks from a text prompt — game soundtracks, ambient backgrounds, prototyping. | ACE-Step 1.5 (Base / Turbo + XL variants) |
Text & Documents
| Module | What it does | Models |
|---|---|---|
| AI Chat | Private chat with the latest open-source language models. No usage caps; nothing sent to a third-party server. | Pick from the Model Center |
| OCR / Document Extraction | Extract text from images and scanned documents — receipts, screenshots, photos of pages. | Nanonets OCR S |
| Embeddings | Generate vector embeddings from text — for retrieval, RAG pipelines, or local semantic search. | BGE M3 |
Model Center
The hub inside the app for browsing, downloading, and switching models. Indicates VRAM fit for your hardware. New models land here regularly as the open-source ecosystem ships them.Next
Hardware requirements
Which modules run on your GPU.
Earning model
What you make per workload, and how to pull GUSD out.