Don’t Believe Us. Prove It Yourself.
Every RAMDeck cluster ships with its own built-in benchmark button. But while you wait, here are raw, unedited proofs of our own machines. For full performance logs and telemetry across all models, read What Speed Looks Like on a Pile of Old Computers →
RAMDeck Can Now Use a Remote GPU Over the Internet (Same Speed, Tested with Tailscale)
Your AI cluster is no longer limited to a single room or Wi-Fi router. Watch a live demonstration of RAMDeck connecting to a remote GPU miles away over the internet via built-in Tailscale—pooling remote VRAM and RAM with local machines at virtually identical token generation speed (31 tok/s vs. 30–32.4 tok/s locally). Read the Full Benchmark & Guide →
Introducing the RAMDeck Node USB: Pool Any Computer’s RAM in Seconds (Zero Install, Unplug & Go!)
Watch how the self-contained RAMDeck Node USB turns any MacBook, Windows PC, or Linux machine into an active cluster node in seconds with zero web downloads—and how one single USB drive can be unplugged to sequentially activate every computer on your LAN, stop cleanly with Ctrl+C leaving zero files behind, or even act as your Primary node. Included FREE ($29 retail value) with every RAMDeck Hub backed on Indiegogo! Read the Full Guide →
RAMDeck Setup Guide: From Unboxing to Chatting in 5 Minutes
Watch the complete end-to-end setup: plug the RAMDeck Hub into your router, pair an RTX 3060 Mini PC with an Apple Silicon Mac mini over local Ethernet, load an 18GB Qwen 30B model in 1 minute 10 seconds, run the one-click benchmark at ~30 tokens/second, and get your first private chat response.
1. When SSD Memory Mapping Beats GPU Sharding (Frontier MoE Models)
We test RAMDeck’s new vertical MMAP against horizontal cluster sharding on MoE models. When the model fits in cluster GPUs, sharding hits ~29 tok/s; when it spills to slow remote CPUs, vertical NVMe MMAP runs over 2x faster (6.89 vs 3.4 tok/s) and loads in under a minute.
2. Live iPhone RAM Contribution, Pause & Resumption
Watch an iPhone dynamically contribute unified RAM and GPU shaders to an active cluster, followed by a live pause and graceful cluster resumption when a node temporarily disconnects.
3. Sharding a 30B Model Across a MacBook & 2 iPhones
An 18GB Qwen 30B model exceeded the RAM of an M3 MacBook. By pairing two iPhones (iPhone 16 + iPhone XS) over local Wi-Fi, RAMDeck pooled their unified memory and Metal 3 GPU shaders to run the model at 28 tok/s (load <30s, turnaround <5s) with zero cloud routing. Download on App Store →
4. Air-Gapped Private RAG Knowledge Base
We proved it on the same 12GB Acer laptop holding the durable backup index. Before ingestion, the chatbot invented a wrong answer about our own product; one FAQ file later, it answered correctly, straight from the document.
5. Coding from Another Room
We loaded a 7B model from the 12GB Acer laptop. RAMDeck routed the inference to the RTX 3060 PC, hitting 43 tok/s. We coded from a Mac Mini — a third machine — using Continue.dev pointed at RAMDeck’s API, successfully summarizing a multi-thousand-line file.
6. Adding an Android Phone to a Live Cluster
We add an Android phone as a genuine fourth contributing node while the cluster is live, showing the automatic rebalance triggered by the new device joining.
7. Loading a 27B Model on an Old Laptop
A 27B parameter model is roughly 16GB on its own — more than the 12GB Windows laptop used as the primary node could hold alone. RAMDeck splits it across four devices with the laptop staying primary.
8. Loading a Massive 40B Model Across 3 Devices
Pushing further: we load a massive 40B parameter model (~25GB) across a 3-device cluster. Full load completes in roughly 2.5 minutes, running at approximately 16 tokens/sec.
9. Running a 13B AI Model Across 3 Home Devices
We split a 13B parameter AI model across three devices already on hand — an old laptop, a mini PC with an RTX 3060, and a Mac mini — running at 11–14 tokens/sec over plain Ethernet.
Take Back Control of Your AI
134% Funded on Indiegogo! Order the RAMDeck Hardware Hub directly for $199 and receive 1 Official RAMDeck Node USB ($29 retail value) completely FREE ($0 Free US Shipping / $20 International · Est. Shipping: January).
Buy RAMDeck Now