Every Monero wallet has to talk to the network through a node. You can run your own or point your wallet at a remote node run by someone else. Both work. They differ in what they leak, how much work they cost, and what happens when things go wrong. This guide lays out the real tradeoffs so you can pick based on your own situation, not on slogans.
What a node actually does for your wallet
Your wallet holds your keys and your view of the balance. The node holds the blockchain and answers the wallet’s questions: what outputs belong to me, what is my unlocked balance, has my transaction been mined. The wallet scans outputs the node supplies against your private view key. That division of labor is the root of every tradeoff in this article, because the node sees your wallet ask those questions.
What a remote node learns about you
A remote node operator can see the IP address connecting to it, when connections happen, and which blocks get scanned. It cannot see your balance, your addresses, or who you pay. Ring signatures, stealth addresses and RingCT hide transaction details from everyone, including the node. What leaks is metadata: this IP, at these times, scanning a wallet active around a given block. We broke down exactly what a remote node operator can and cannot observe in our earlier article on what remote node operators see. The short version: for most people it is a modest exposure, and it matters more if someone can watch network traffic and correlate timing.
Useful middle ground: a remote node over Tor. Both the official GUI and Feather support connecting through Tor, which strips the IP from what the operator sees. You keep the convenience of a remote node and remove its biggest leak. The cost is slower sync and occasional connection flakiness.
What running your own node gets you
Your own node means your wallet asks its questions of a machine you control. Nothing about your scanning leaves your network. Two practical routes exist:
A full node stores and verifies the whole chain. Right now that is roughly 200 GB of disk, growing slowly, and the initial sync takes several hours to a couple of days depending on hardware and bandwidth. A pruned node keeps the same verification guarantees while storing about a third of that, which is what most home setups should use.
A node is not the same as mining. Running a node just means keeping a copy of the ledger and serving your own wallet. It needs a few GB of RAM and modest CPU after sync. Mining is a different activity with different hardware economics.
The chain is over 3 million blocks and your node verifies every one of them during the initial sync. A cheap VPS with 40 GB of disk handles a pruned node comfortably. Verify the daemon download before you run it, the same way you would verify your GUI download.
The honest costs of self-hosting
Initial sync is the main hurdle. On a decent home connection expect 6 to 24 hours for a pruned sync, longer for full. After that the node stays current with minimal bandwidth. The other cost is operational: the daemon occasionally needs a restart after version upgrades, and a node that is off when you want to transact means waiting for it to catch up. If you transact rarely, start the node a few minutes before you need it and accept the catch-up time.
One real risk deserves stating. A node you control but expose badly is worse than a good remote node. Do not open the daemon port to the internet with unrestricted RPC. Keep it bound to localhost, or if other devices need it, use the restricted RPC port and a strong password on the unrestricted one.
Practical recommendations by situation
Checking a balance weekly, small amounts: a remote node over Tor. The setup cost of a personal node buys you little here.
Regular transactions, moderate amounts: run a pruned node on a spare machine or VPS. The sync is a one-time cost and every future transaction gets the full benefit.
Mobile wallet: remote node by default, that is what the official Monerujo and Cake Wallet apps do, and both let you add Tor. Pair it with a desktop node for larger amounts.
Long-term holder: a node you control, paired with a tested restore from seed. The combination means your access to your own funds depends on nothing and nobody else.
Common misconceptions
Running your own node does not make your transactions more private on the network side. Dandelion++ and the decoy selection work the same way regardless of whose node you use. What changes is who can observe your wallet’s activity pattern. Likewise, using a remote node does not weaken the cryptography one bit. The privacy loss is about the observer, not the math.
One more distinction people miss: a remote node does not hold your keys at any point. Your wallet downloads the outputs and does the scanning locally, on your device. Remote node, own node, the private keys never leave your machine. The choice is only about who serves the blockchain data to the wallet doing that scan, which is exactly why the stakes are metadata and not funds.
The short version
Remote node: zero setup, leaks timing and IP to the operator, fixable with Tor. Own pruned node: roughly 70 GB of disk and one long sync, then nobody watches your wallet at all. For casual use, a Tor remote node is a reasonable choice. For a wallet that matters, spend the weekend on your own node and be done with the question.