🐝About Sogni Fast Worker

#How Sogni Fast Worker Works

Sogni Fast Worker is a Docker-based application that runs in the background on your Windows or Linux machine. It’s designed to be "plug-and-play," so you can start contributing GPU inference tasks on (and earning from) the Sogni Supernet render network in minutes.

• Run 24/7 on dedicated hardware
• Deploy on cloud compute platforms (like Salad.com) that accept Docker images
• Run part-time on a personal gaming PC when idle

#How Do You Earn With Sogni Fast Worker?

Once your Fast Worker is running, it automatically downloads supported models and waits for creative AI tasks on the Sogni Supernet. By default, regular jobs paid with SOGNI token or Spark Points earn SOGNI for your worker each time it completes a job. Those earnings accrue in the worker's associated Sogni Account Smart Contract, with real-time accounting and claiming available at dashboard.sogni.ai.

Unlimited subscription-covered jobs are a new optional second income stream. Workers who opt in can receive steady, high-traffic subscription jobs that accrue to the monthly Worker Subscription Revenue Share Pool instead of settling per job in real time. Use the upcoming dashboard.sogni.ai/subscription-earnings page to review terms, opt in or out, and track monthly pool earnings.

#How Much Can I Earn With Sogni Fast Worker?

As a rough estimate, per our historical Daily Fast Worker Leaderboard from Testnet 4, a high performing Nvidia RTX 5090 machine may complete upwards of 10,000 image inference jobs every 24 hours. Factoring NET proceeds of $0.0025 USD per inference job, inclusive of network fees, you may earn up to ~$25 on a peak day (roughly $350–$750 per month depending on demand). Video, Audio, and LLM jobs are also available on the network.

*Your mileage may vary, this is an estimate based on expected peak demand and network fees.

#Prerequisites to Run a Fast Worker

1. A free Sogni account (sign up at app.sogni.ai or use an existing one from Mac/iOS apps).
2. Docker Desktop for Windows or Docker Engine for Linux. Both are free.
3. A supported Nvidia GPU. The list of supported GPUs are generally any recent Nvidia GPU with at least 16GB of VRAM. This list currently includes:
   RTX 5090 - Ideal for Image and most Video inference
   RTX 4090 - Ideal for most Image Inference and some video
   H100 - Ideal for all Video Inference
   RTX Pro 6000 - Ideal for all Video Inference
   RTX 5080
   RTX 4080 Super
   RTX 4080
   RTX 4070 Ti Super
   RTX 5070 Ti
   RTX 3090 Ti
   RTX 3090
   RTX 6000 Ada Generation
   RTX 5000 Ada Generation
   RTX A6000
   RTX A5000
   L40S
   H200
   A100
   B200
   Tesla V100
   RTX Pro 4000
   RTX Pro 4500
   RTX Pro 5000
4. A system with:
   RAM: At least 40GB for bare-metal installs, 30GB for remote Docker Container hosting. Ideally 48GB+ allocated for Video / ComfyUI Worker.
   Disk Space: For Sogni Comfy Worker automatic downloads, budget about 470 GB for RTX 5090 and higher-VRAM GPUs (32GB+ VRAM), or about 435 GB for 24GB GPUs like RTX 4090 and RTX 3090. These July 2026 figures cover the current Comfy Worker model cache; reserve additional space for Docker images, temporary downloads, logs, and the operating system. Stable Diffusion Worker model storage is optional and depends on how many models you host.
   Internet: Min 5-10 Mbps recommended. Internet speed impacts job priority.
5. Sogni API Key + Sogni Fast Worker NFT (free). When you start the worker (worker-start.sh or worker-start.bat), you will be guided through a one-time worker authentication process. Please check the Fast Worker FAQ page for the latest information on acquiring the Fast Worker NFT.

There are currently three active versions of the Sogni Fast Worker. The former Sogni Flux Worker is retired and has been replaced with the Sogni Comfy Worker.

1. Sogni Comfy Worker
   Our latest worker type for the newest open-source image, video, and music workflows. It supports LTX-2.3 Video with audio, Wan 2.2 video models, and current frontier image workflows including Krea 2 Turbo, Z-Image / Z-Image Turbo, Chroma, Qwen Image Edit 2511 Lightning, Qwen Image 2512, Flux.2 Dev, FLUX.1 Krea Dev, and FLUX.1 Schnell. It also supports music generation with Ace-Step 1.5 XL, the latest Ace-Step 1.5 family, plus non-XL Ace-Step 1.5 workflows. Krea 2 Turbo is the newest open-source Krea image model on Sogni and is separate from Flux.1 Krea. This is the replacement for the retired Sogni Flux Worker.
2. Sogni Stable Diffusion Worker
   For running any of the 95+ Stable Diffusion community models plus advanced Controlnets.
3. Sogni LLM Worker [NEW: March 2026]
   Sogni's powerful new LLM worker runs the latest Qwen3.6 35B A3B model with vision support. It supports both regular abliterated model versions (abliterated currently supports Qwen3.5).

Retired: The Sogni Flux Worker is retired and replaced with the Sogni Comfy Worker. Its Flux and frontier image models have moved to Comfy Worker; do not deploy new Flux Worker instances.

Which Worker Should You Choose?

• Stable Diffusion Worker: If you want access to a wide variety of models (95+) and advanced Controlnets. It does require more disk space (250gb+) if you choose to host every model but that is optional. If you have a supported GPU with only 16gb of ram you'll get more performance and more jobs with this model.
• Comfy Worker: Choose this for Krea 2 Turbo, Z-Image Turbo, Chroma, Qwen Image Edit 2511 Lightning, Flux-family workflows, Wan/LTX video, Ace-Step 1.5 XL music, and the newest Comfy-hosted workflows. It replaces the retired Sogni Flux Worker. Video workflows are best for 24gb VRAM+ GPUs such as RTX 5090, H100, and RTX Pro 6000, but many image and music workflows can run on 16gb VRAM or more. As of July 2026, hosting every automatically supported Comfy workflow requires about 470 GB of model cache on RTX 5090 and higher-VRAM GPUs (32GB+ VRAM), or about 435 GB on 24GB GPUs like RTX 4090. You may select specific workflows to reduce disk use.
• Sogni LLM Worker: Our latest worker type carries the most advanced LLM models that can fit on 24gb of vram or more. Jobs include text chat, vision (image scene analysis, OCR), and powering Sogni Tools / Sogni Intelligence.

You can switch modes at any time by rerunning the installer.

#Leaderboard Rewards

Regular Spark- and SOGNI-paid jobs remain the default earning path and continue to settle to your worker in real time for each completed job. The Worker Subscription Revenue Share Pool is optional: it adds subscription-covered jobs as a second income stream for workers who accept the Subscription Earnings terms and enable participation from the upcoming Subscription Earnings dashboard.

For full pool mechanics, attribution, and payout details, see Worker revenue share.

#Pending Updates

We’re always actively developing additional features for Sogni Fast Worker. On the roadmap: model training, text-to-3d, upscaler jobs.

#Legal and Liability Disclaimers

By running this worker application, you agree to the Sogni Terms of Service & Privacy Policy and the Leaderboard rewards Terms and Conditions.

Compliance with Local Laws

By participating, you agree to comply with all relevant laws. Sogni is not responsible for legal issues. Sogni monitors regulations to ensure compliance.

Assumption of Risk

Operating a node involves risks like hardware wear, electricity cost, and software issues. By participating, you accept these risks.

Limitation of Liability

Sogni is not liable for direct or indirect damages, financial loss, or device damage, even if advised of such possibilities.

No Guarantee of Earnings

Earnings vary; no minimum is guaranteed.

No Employee or Agency Relationship

You’re an independent participant, not an employee of Sogni.

Indemnification

You agree to hold Sogni harmless from any claims arising from your participation.

Software Use at Your Own Risk

The software is “as is” without warranties. Use at your own risk.

Agreement to Terms

By operating a node, you acknowledge and agree to these terms and disclaimers.

Need help? Join our Discord!

If you have any issues or questions setting up your Sogni Fast Worker you can reach out on Discord or via email to [email protected] for technical support.

Join Sogni Discord ✨