RpcNode Toolkit

Control plane for your infrastructure

Run full nodes
without the SSH routine

A self-hosted control plane for your node fleet. Connect a server, choose a network, and operate full-history nodes from one interface — with your disks, keys, and RPC endpoints.

Download releasesRead the docsExplore the catalog

Free and self-hosted · Ready JARs on GitHub Releases · Supported and recommended: Ubuntu 24.04 LTS (amd64 or arm64). Other operating systems have not been tested.

Toolkit admin setup screen for creating an administrator account
Admin setup
21networks
53environments
fullhistory, not prune
1agent per host
$0Toolkit license

Free

No license. No seat fee.

RpcNode Toolkit is free software. You do not pay us to run the panel, the host agent, or as many full nodes as your hardware can hold. There is no Toolkit subscription and no per-node charge. You pay only for the server, disk, and bandwidth you already own.

Self-host

On your server. Under your keys.

Install it on Ubuntu 24.04 LTS you control — a dedicated box, a VPS, or a rack at home. The control panel and the node processes stay on your infrastructure. RpcNode does not hold the chain, the RPC port, or the agent token. Your disks, your firewall, your RPC.

Operate

You run the node. You change the config.

After install the panel is the control plane: Node Config edits the live files (bitcoin.conf, java-tron, xrpld, geth flags), restart, client pin, logs, public RPC. Install choices — TRON snapshot flavor, XRPL history window, disk layout — are saved on the node. No SSH to “fix conf by hand”.

What Toolkit is

Toolkit is a self-hosted control plane for blockchain full nodes. One Ubuntu 24.04 LTS agent per server, one panel for the fleet. You add a host, pick a network — Bitcoin, Ethereum, TRON, Solana, and 17 others — and the panel walks install, snapshot when the chain has one, start, and sync. After that you administer the node from the same UI: Node Config, logs, disk IOPS, client updates, restart, public RPC.

It is not a hosted RPC product and not a light wallet. Every node is a full history node: no prune-by-default, no seven-day window. Bitcoin ships with txindex=1 and prune=0. EVM chains run unpruned / archive as the network expects. Stellar keeps history instead of the stock week. Synced means a live tip and history proof — not a green badge on an empty ledger. I am Toolkit. I put the full node on your server.

Pre-built JARs ship on GitHub Releases: rpcnode-server.jar, rpcnode-agent.jar, and rpcnode-cdn.jar. Download those if you do not want to compile from source.

The installer is a single script. It places rpcnode-api-agent and rpcnode-system-agent on the host, enables systemd, and prints an Agent URL plus AGENT_API_TOKEN. You paste those into the panel. There is no remote SSH from our cloud into your box.

Chains that bootstrap from archives can pull from a snapshot CDN. Run your own with rpcnode-cdn.jar on a host you control, or use the public catalog at cdn.rpcnode.dev.

  1. 01

    Panel on your ops host

    Run the self-hosted panel (Docker on a machine you operate). It talks to agents. It is not the fullnode.

  2. 02

    Agent on each node server

    One curl per Ubuntu 24.04 LTS host. That agent is the identity of the server — Bitcoin, TRON, and Solana are profiles on the same product, not separate installers.

  3. 03

    Add node in the UI

    Network → environment → server, then the NODE SETUP wizard: Check ports → Install → Snapshot (when the chain has one) → Start / sync → Healthy. Documentation covers the current setup. Sync %, logs, per-disk IOPS, and RPC stay on the node page.

What you get on the node

  • Network state — Sync status on the node page: SYNCED or catching up, your slot / height vs the network tip, peers, Healthy / Running / Node online. The bar is honest (IBD, snapshot, slots behind) — 100% only with live tip and history proof, not a green badge on an empty ledger.
  • Host & node consumption — CPU, RAM, net ↓/↑, disk IOPS side by side: the whole server vs this node unit (cgroup). Disk tabs per NVMe: read/write IOPS, MB/s, load %, free space. Solana ledger vs accounts are not one blended number.
  • Journal and agent logs in the panel. Host audit for discrete ops, not process chatter.
  • Node Config — read and edit the node’s live config from the panel, then restart. Install options (which snapshot, how much XRPL history) stay on the node record.
  • Client pin from the catalog. Re-apply a new version from the node page.
  • RPC load — Fullnode Go RPC (public proxy): RPS 1m / 5m, latency p50 / p95, in-flight, 4xx / 5xx / 502 / 503 / upstream errors, and a live RPS chart. Clients hit the proxy, not the node process directly.
  • Telegram alerts for disk, CPU, RPC errors, node down / recovered.
Is RpcNode Toolkit free?
Yes. The panel, the host agent, and node administration are free to self-host. No Toolkit license. You cover your own servers and traffic.
What does self-host mean here?
You install the software on hardware you control. The fullnode binary, datadir, and RPC listen on that machine. We do not run the node for you and we do not keep a copy of the chain.
Can I change the node config after install?
Yes. Open the node page → Node Config. The panel loads the files the agent wrote (conf, toml, HOCON, flags) and lets you edit and apply them. Install-time choices such as a TRON snapshot flavor are stored on the node; switching a snapshot kind means wipe and Add node again — the LevelDB is already a different dataset.
What does network state show?
Sync status: your node vs the public tip (slot, height, block), peer count, and badges (Synced, Healthy, Running). Synced is not “RPC answers” — the agent proves history, then paints 100%.
Can I see RPC load on the node?
Yes. Fullnode Go RPC on the node page is the public proxy: requests per second, latency, in-flight, and error classes (5xx, 502, 503, upstream). Zero RPS means no clients are calling that endpoint yet — the node can still be Synced.
Does Toolkit monitor host and node consumption?
Yes. Host & node splits the server from this fullnode unit: CPU, RAM, network Mbps, disk IOPS (cgroup accounting). Each physical disk has its own tab — IOPS, load %, free space. Telegram can alert on disk and CPU.
Can I put it on my own VPS or bare metal?
Yes. One agent per Ubuntu 24.04 LTS server; several networks can share a host on separate ports (some chains allow only one environment per machine). You need enough disk for full history. Supported and recommended: Ubuntu 24.04 LTS (amd64 or arm64). Other operating systems have not been tested.
Which OS do you support?
Supported and recommended: Ubuntu 24.04 LTS (amd64 or arm64). Other operating systems have not been tested.
How is this different from managed RPC on rpcnode.dev?
Managed RPC is our hosted endpoints. Toolkit is the open install path if you want the same catalog of full nodes on your servers. You can use one, the other, or both.
Can I run my own snapshot CDN?
Yes. Toolkit ships rpcnode-cdn.jar so you can mirror snapshot archives on your own disk and serve them to your fleet. We also publish a public CDN at cdn.rpcnode.dev.
Where do I download pre-built binaries?
From github.com/rpcnode/toolkit/releases. Each tagged release attaches rpcnode-server.jar, rpcnode-agent.jar, and rpcnode-cdn.jar (plus checksums). You can still build from source if you prefer.
Can you add another network? Other questions?
Join the Telegram group, write to admin@rpcnode.dev, or open rpcnode.dev — new chains, install problems, or anything else about Toolkit.

DocumentationGitHub ReleasesSupported networksSnapshot CDNcdn.rpcnode.devTelegramSource on GitHubadmin@rpcnode.devrpcnode.dev

Community

Ask in Telegram

Questions, ideas, a network you want added — write in @rpcnodetoolkit. The group is open.

Open Telegram

Networks