Source: https://docs.sogni.ai/run-a-worker/fast-worker/running-on-nosana/ # Run a Sogni Worker on Nosana Nosana supplies the rented NVIDIA GPU. Sogni Comfy Worker connects it to the Supernet, receives eligible AI jobs, and credits rewards for completed work. You manage the rental in Nosana and the worker in the [Sogni Workers dashboard](https://dashboard.sogni.ai/fast-workers/). As of September 2026, the **Sogni Comfy Worker** template is proposed in [Nosana template PR #172](https://github.com/nosana-ci/pipeline-templates/pull/172) and is not yet available in the public catalog. Once it appears in Nosana Deploy, use the steps below. Its README also contains a complete setup guide. This template runs the Fast Worker service; it does not open a standalone ComfyUI editor. ## [#](https://docs.sogni.ai/run-a-worker/fast-worker/running-on-nosana/#1-prepare-your-two-sogni-values)1\. Prepare your two Sogni values | Value | What to do | | --- | --- | | **Sogni API key** | Sign in at [dashboard.sogni.ai/api-key](https://dashboard.sogni.ai/api-key) and copy your account API key. | | **Fast Worker NFT token ID** | Open [nft.sogni.ai](https://nft.sogni.ai/) with the wallet associated with your Sogni account. Use a Fast Worker NFT you own and copy its numeric token ID. | One GPU worker needs one unique NFT. Stop any other worker using that NFT before deploying here. An API key can be shared by workers on the same Sogni account; the NFT token ID must be different for each simultaneously running worker. Keep your deployment confidential. Do not publish a configured job definition or upload it to public IPFS with your API key inside. Follow [Nosana's confidential-job guidance](https://learn.nosana.com/deployments/jobs/job-definition/confidential.html), and never share credentials in screenshots or support messages. ## [#](https://docs.sogni.ai/run-a-worker/fast-worker/running-on-nosana/#2-choose-a-gpu)2\. Choose a GPU Create or sign in to your account at [Nosana Deploy](https://deploy.nosana.com/) and fund it for the rental. Select **Create Deployment**, choose **Sogni Comfy Worker**, and select the matching preset: | Preset | Suitable GPU examples | Recommended host RAM | Disk capacity to plan for | | --- | --- | --- | --- | | **24 GB GPUs** | RTX 3090 or RTX 4090 | 64 GB | 1 TB | | **32 GB+ GPUs** | RTX 5090 or RTX PRO 6000 | 64 GB; 128 GB for a 96 GB GPU | 500 GB | Both presets require an x86-64 Linux host with **NVIDIA driver R580 or newer**. A GPU's name alone does not establish its driver or the memory and disk available to your container. Setup checks for at least **47 GiB of RAM visible to its container** and rejects smaller hosts before the worker starts. It does not enforce the driver or free disk requirements; confirm the market's specifications and inspect the assigned host. A RAM rejection can cause a retry or delay, and does not reserve a suitable node. The 24 GB preset preloads the general image, video, and audio catalog for that GPU size. The larger-GPU preset uses MiniMax video/music and FlashVSR upscaling, matching Sogni's larger-GPU Nosana catalog. The worker checks which workflows it can actually serve on its hardware with the files available. As of September 2026, the model files total about **529 GB** for the 24 GB preset and **116 GB** for the larger-GPU preset. These figures change with the catalog. The disk recommendations include extra room, but are not a guarantee of free space on an assigned host. Container images, temporary files, caches, and future models also need space. ## [#](https://docs.sogni.ai/run-a-worker/fast-worker/running-on-nosana/#3-enter-your-values-and-deploy)3\. Enter your values and deploy 1. Choose the GPU market and keep **Replicas = 1**. 2. Use the **Infinite** deployment strategy for an ongoing worker. 3. In the job definition form, find **sogni-worker** and expand **Environment variables**. 4. Replace `REPLACE_WITH_YOUR_SOGNI_API_KEY` in **API\_KEY** with your Sogni API key. 5. Replace `REPLACE_WITH_YOUR_FAST_WORKER_NFT_ID` in **NFT\_TOKEN\_ID** with your numeric token ID. 6. Leave the resource loader, image, model mounts, and health endpoint at their defaults. Review confidentiality and the rental cost, then deploy. To run a second GPU, create a second one-replica deployment with another NFT. Raising the replica count copies the same NFT into every worker and makes them compete for one identity. ## [#](https://docs.sogni.ai/run-a-worker/fast-worker/running-on-nosana/#4-wait-for-setup-and-verify-the-worker)4\. Wait for setup and verify the worker The template pins the Sogni-owned setup image, [Nosana Worker Configurator 0.1.0](https://hub.docker.com/r/sogni/nosana-worker-configurator/tags?name=0.1.0), by its immutable digest. It checks available RAM, resolves the recommended Sogni worker image, and prepares the model resources. A cold host may need tens of minutes or longer to download them. Watch the progress in Nosana; repeated restarts can delay setup or move the job to another cold host. Open [Sogni Workers](https://dashboard.sogni.ai/fast-workers/) and select your NFT. Confirm its online status, reported version, GPU, and available models. Review **Worker Health** for startup progress and **Job History** as work arrives. Nosana showing a running container is not proof of a completed Sogni job. The template's port **8001** is a dedicated health endpoint. `/startup` passes when startup is progressing normally, while `/readiness` passes when the worker can accept work. No public ComfyUI or worker control port is needed. ## [#](https://docs.sogni.ai/run-a-worker/fast-worker/running-on-nosana/#settings-updates-and-storage)Settings, updates, and storage Open the worker's **Settings** tab in Sogni for supported operator controls. See [Using the Workers Dashboard](https://docs.sogni.ai/run-a-worker/fast-worker/worker-dashboard/) and [Advanced Configuration](https://docs.sogni.ai/run-a-worker/fast-worker/sogni-fast-worker-advanced-configuration/). Preferences cannot add model files to Nosana's preloaded resource bundle; retain the managed download defaults unless following a specific Sogni recommendation. Each new Nosana job resolves Sogni's recommended, versioned worker image. A running job stays on its existing image until replaced. Read the [Comfy Worker release notes](https://docs.sogni.ai/run-a-worker/fast-worker/release-notes/comfy-worker/) before an intentional update and allow for queuing and another download after replacement. Model files use Nosana's remote-resource cache under `/data-models`; reuse depends on the assigned host. This template does not add a portable persistent `/data` volume. Replacement can lose local worker state. Compatible dashboard settings have best-effort recovery through Sogni, but this is not a backup of the worker's entire data directory. ## [#](https://docs.sogni.ai/run-a-worker/fast-worker/running-on-nosana/#if-something-goes-wrong)If something goes wrong | Symptom | Next step | | --- | --- | | Queued | Wait for a host or choose another compatible market. | | Download progress is advancing | Let setup continue. A cold cache takes longer. | | RAM, storage, or driver error | Choose a host meeting the preset's requirements. Restarting on the same unsuitable host will not fix its hardware. | | Resource JSON error during setup | Read the **resources** operation's logs first. A preceding RAM or configuration error can also produce a downstream resource-parse error. | | Running in Nosana but offline in Sogni | Check both entered values, NFT ownership, and whether another worker is using that NFT. Review startup logs without sharing credentials. | | Online but no jobs | Check available models and Worker Health. Job arrivals depend on demand and eligibility. | Nosana rental charges are separate from Sogni rewards. Neither job volume nor cost recovery is guaranteed. Stop the deployment in Nosana when you want to stop renting the GPU; closing either dashboard does not stop the rental. For help, see the [Fast Worker FAQ](https://docs.sogni.ai/run-a-worker/fast-worker/fast-worker-faq/), email [app@sogni.ai](mailto:app@sogni.ai), or join [Sogni Discord](https://discord.com/invite/2JjzA2zrrc).