Do my files pass through or get saved on any server?

No. Nearsend is strictly a peer-to-peer (P2P) platform. The signaling server is only used to exchange temporary session tokens and WebRTC handshake coordinates. Once connected, your files, text, and clipboard stream directly between your devices via WebRTC DataChannels in 64KB encrypted chunks.

How does the Voice Pairing work securely?

When you generate a voice pairing phrase, the host device creates an ephemeral 3-minute token. When spoken on the guest device, browser speech recognition converts the acoustic phrase into the token. The server validates the match and immediately triggers an explicit confirmation prompt on the host device before any connection is allowed.

Does Nearsend work across different operating systems?

Yes! Nearsend works seamlessly across any modern web browser — Windows, macOS, Linux, Android, iOS, iPadOS, and ChromeOS. No native apps or software installations are required.

Is there a file size limit on Nearsend?

Because transfers stream directly via WebRTC without hitting a central server memory limit, you can send files ranging from kilobytes to multi-gigabyte videos. Nearsend incorporates backpressure flow control to ensure smooth transmission without exhausting device RAM.

Does Sending a File Ever Touch a Server? A Plain-English Breakdown

September 13, 2026

"Peer-to-peer" and "no server involved" get thrown around a lot in file-sharing marketing, and not always accurately. It's worth walking through what actually happens, step by step, in a real browser-based peer-to-peer transfer — because the honest answer is more specific than a flat yes or no.

Step 1: finding each other

Two devices that have never met can't magically connect. They first exchange a small handshake through a signaling server — tiny setup messages, not your file. This is the step a QR code, PIN, or voice phrase kicks off: it's how the two devices agree they want to talk to each other.

Step 2: working out the network path

Once the two devices know about each other, they try to work out the most direct route between them, using STUN to look past whatever router or NAT each one sits behind. In the common case, this succeeds, and a direct connection opens between the two devices — no server on the data path at all from this point on.

Step 3 (sometimes): a relay, not a stop

Certain networks block a direct path outright — some mobile carriers, some locked-down corporate firewalls. When that happens, a TURN server relays the encrypted stream between the two devices in real time. It's forwarding traffic it can't read, the same way a sealed envelope passing through a mail sorting facility isn't opened along the way — it isn't storing your file.

What never touches a server

On the common, direct-connection path: the bytes of your file, its filename and contents, none of it is ever written to a server's disk. It exists on the sending device, in transit across an encrypted connection, and on the receiving device — that's the whole path.

So is 'no server involved' accurate?

Not quite as an absolute claim, and it's worth being precise about that: a signaling server is genuinely required for the two devices to find each other, and a relay is sometimes required to get the data across a restrictive network. What's true is narrower and more useful — your file's actual contents are never stored on a server, whether the path ends up direct or relayed. That's a real, verifiable distinction from a tool that uploads your file to cloud storage first and hands out a download link second.