Run your own InfiniteOcean node. All compute is free — search, SQL, functions, encoding. No limits, no meters. Your hardware, your rules.
We'll email you when your node is ready to download.
One installer for macOS or Windows. Nothing else to install — just open and go.
Choose how much data stays local. All compute is free regardless of tier. Change sizes anytime.
Your node boots, auto-registers, and serves the full API. Data stays on your machine.
Pick the storage you need. All compute is free on every tier — the only difference is how much data stays instant on your node.
| Tier | Local storage | Compute |
|---|---|---|
| 1gb | 1 GB | Unlimited |
| 10gb | 10 GB | Unlimited |
| 100gb | 100 GB | Unlimited |
| 1tb | 1 TB | Unlimited |
| custom | 2 TB+ | Unlimited |
Your data is instant because it's local. When your node runs out of space, new data can't be stored locally anymore — it loses the speed advantage. Delete data or pick a bigger tier to keep everything instant.
No time limits on functions. No compute meters. Search, SQL, encoding — all free on your hardware.
Drops, SQL, text + vector search, streams, webhooks, functions, media storage. Every feature. Nothing held back.
No license key, no env vars, no signup. Boot and go. Auto-registers and starts serving immediately.
Built-in Python runner. Execute functions with full ocean.sql(), ocean.search(), ocean.write() access. No time limits.
Central server unreachable? Node starts standalone. Registers when connectivity returns. No downtime.
All data persisted locally. Survives restarts, upgrades, redeployments.
Submit a video or audio URL — encoding runs locally on your hardware. Output is HLS, streamed through the API. Free.
Your node caches reads from the network. Search, query, entity lookups, and exports are served from the nearest node — lower latency for your users.
Link two or more nodes together. Local data automatically replicates between them for redundancy. If one goes down, the other keeps serving.
Write drops with local: true — they never leave your node. Not forwarded, not synced, not cached anywhere else. Your data, your machine, period.
Every node contributes a small slice of compute to the network. Search, SQL, and functions run across all nodes automatically — no configuration needed.
Up to 10% of your compute helps other users. The rest is yours. All compute is free — for you and the network. The more nodes, the faster everything gets.
Link nodes on your network so local data stays replicated between them. If one goes offline, the other keeps serving.
Open the app on both machines. Each shows a Your Node URL like http://192.168.1.42:4973 on the dashboard.
Copy the Node URL from one machine and paste it in the Peers section of the other. Do the same in reverse.
Data written with local: true replicates between peers every 30 seconds. Non-local data syncs to central as usual.
Peer sync is bidirectional. Node A adds Node B, and Node B adds Node A. Adding a peer restarts the container briefly (~2 seconds).
Got a Linux VPS with a public IP? One command will turn it into a TLS-enabled InfiniteOcean edge node with peer sync.
Installs everything needed automatically. Works on Debian, Ubuntu, CentOS, Fedora.
Point a domain at your server and the script sets up automatic HTTPS certificates. Works with IP-only too.
Starts the node and registers it as a public peer with bidirectional sync. Fully automatic.
Linux server (any distro with apt, yum, or dnf), root access, and a public IP. A domain is optional — works with IP-only too (HTTP mode). Port 80 (and 443 if using a domain) must be open.
| Endpoint | Description |
|---|---|
| GET /node/status | Tier, storage usage, capacity |
| GET /whoami | Account info, capacity summary |
| POST /node/claim | Link node to your account with X-Ocean-Key |