Ocean Orchestrator connects VS Code, Cursor, Windsurf, and Antigravity to pay-per-use GPU compute on Ocean Network. One click to run. Results back locally.
Selected the resources that fit your run
Browse live GPU environments, filter by spec and region, and lock your resources in one click
Open Ocean Orchestrator in VS Code, Cursor, Windsurf, or Antigravity. Submit the job and monitor live logs without leaving your editor
Results, logs, and metrics are saved automatically to your local folder the moment the job completes
GPU compute workflows often slow you down before the job even starts. Comparing options, cloud setup, switching dashboards, and paying for idle infrastructure kills momentum.
You want a direct path: choose your GPUs, run a job, pay only for execution time.
No forced bundles, choose exactly what your workload needs.
Funds are held securely and released only after your job completes.
Works with VS Code, Cursor, Windsurf, and Antigravity.
Launch with one click and track progress live in your editor.
Results and logs are saved automatically — no manual retrieval.
Budget is set before the job starts. Funds held securely in the escrow contract.
Your containerized workload runs on the selected node. Progress tracked live.
Completion is verified before any payment is processed.
Exact runtime cost released. Any remaining balance returns to your wallet.
Run inference over large input sets and pull back clean outputs and logs.
Turn raw data into artifacts your app can ship with.
Rerun the same job with different resources and iterate faster.
Clear answers so you can run AI jobs with just one click
The quickest way to run your first compute job is to use the Ocean Network dashboard: select the resources you need, then push them to Ocean Orchestrator (you’ll be prompted to install it if you haven’t). Ocean Orchestrator works in VS Code, Cursor, Antigravity, and Windsurf.If you prefer CLI, you can use ocean-cli to submit the job and pull results when processing is complete.
You can run containerized compute jobs like embeddings, model inference jobs, data cleanup, batch processing, and fine-tune model workloads that finish within the job window and produce outputs you can download.
Ocean Nodes are designed to manage failures locally, keeping compute job execution predictable and controlled. If a node goes down mid-run, the job can be restarted on the same node once it becomes available again. Funds are released from escrow only when the node explicitly marks a job as successful.Rerouting is handled by the user, in line with the Ocean Network ethos of giving users full control over which resources are used.This differs when a failure is caused by the algorithm itself. In that case, the job is treated as unsuccessful because the execution failed, not because the node was unavailable. See the dedicated FAQ for algorithm failures.
No. This is no server setup and serverless GPU compute: you choose a preferred Ocean Node with the resources you need, submit a containerized job, and get results back without managing servers or infrastructure
Yes, we have 3 types of free compute. Test CPU compute enviroemnts offered by Ocean Network, Complimentary Credits which gives you access to GPU compute and lets you use resources for $100 USD. The third type of test compute enviroments is the one that can be offered by node operators for the purpose of showcasing their resources for users to experiment with.
Ocean Network does not auto-assign your job to random machines. You pick the node or compute environment you want based on resource price limits and current availability, then submit the job to that specific resource. If it is busy, your job waits until that node becomes available. The dashboard shows availability and max duration per environment, so you can choose predictable compute up front
Ocean Network is a decentralized, peer-to-peer (P2P) compute network for pay-per-use compute jobs that turns idle or underutilized GPUs into usable distributed compute resources.
