FAQ

RAMDeck FAQ — Pricing, Setup, Speed & Device Support

Frequently Asked Questions

Q: Where can I find and order RAMDeck, how much does it sell for, and how do I set it up?

Following our 134% funded Indiegogo campaign ($13,422+ raised), you can order the RAMDeck Hardware Hub by TradeMAV (the TradeMAV RAM pooling AI hub) directly at ramdeck.com for $199 USD—which includes 1 Free Official RAMDeck Node USB ($29 retail value), $0 Free US Shipping ($20 flat international shipping), and estimated shipping in January.

Setup takes under 5 minutes with zero coding: (1) Plug the pre-flashed RAMDeck Hardware Hub into your router via Ethernet and USB-C power; (2) Connect your Windows PCs, Macs, Linux machines, iPhones, or Android phones using either the included zero-install Node USB or our free 1-click apps at ramdeck.com/downloads; (3) Open the browser dashboard to load 30B, 70B, or 120B models across your local network or built-in Tailscale.

Q: What about context size? What happens after a 30-minute conversation?

With local AI, context size (how much of the chat history the model remembers) eats up RAM fast. On a single, memory-constrained machine, a 30-minute conversation usually means the model runs out of context space and starts “forgetting” the beginning of the chat, or simply crashes.

Because RAMDeck pools the memory of multiple machines, your available context limit scales up with your hardware. If you plug an old 16GB laptop into your cluster, that extra pooled RAM can be dynamically dedicated to expanding your context window. A long conversation that would have choked a single machine can now fit comfortably across the cluster, keeping the model sharp and fully aware of the entire session history.

Q: Won’t using my network as RAM be incredibly slow? A gigabit connection (128MB/s) is slower than a cheap SSD swap file!

Your math on gigabit is right, but it’s not really doing what you’re picturing. Where you’re totally right is actual OS-level swap, where a model doesn’t fit and pages get shuffled to disk constantly — that genuinely is bandwidth-bound and it thrashes hard.

But RAMDeck isn’t swapping. It doesn’t mail model weights back and forth every time someone needs to read something. It hands out different “chapters” (layers) of the model to different machines once, at the start. Each machine keeps its chapters in its own RAM for the whole session.

When generating a response, the machines just pass a short summary note (the activation state) down the relay line to each other. What crosses the network is just that short note, not the weights themselves. Real measurements put it around 650KB per token. Even at a decent generation speed (10 tokens/second), that’s only ~6.5MB/sec — nowhere close to a gigabit network’s ceiling. The real bottleneck is the latency of handing that note to the next machine (which is why wired ethernet beats Wi-Fi for RAMDeck), not the bandwidth.

Q: Since distributing chapters at load time takes 50-60 seconds over the network, wouldn’t it be faster to just keep duplicate copies of the model saved on every device’s local SSD?

You are absolutely correct that loading from local SSDs on every machine would eliminate that initial network transfer time. However, it comes with a massive tradeoff: storage bloat and sync burden.

If we kept a local copy everywhere, every device on your network would need enough free disk space to hold the entire model, not just its assigned piece. If you want to run a 30GB model across a 4-node cluster, you’d be burning 120GB of total storage across your fleet. Furthermore, every time you want to try a new model or update an existing one, you’d have to manually copy it to every single machine again.

By using a single-source-of-truth architecture, the primary machine holds the model on its disk, and RAMDeck pushes the required pieces to the cluster’s RAM only when the model starts up. Yes, you wait ~60 seconds for the network transfer when booting a massive model, but you save massive amounts of local storage across your devices and never have to deal with out-of-sync nodes. And remember, that transfer only happens once at startup—during actual generation, the weights stay completely locked in RAM.

Q: Does any of my data leave my network?

No. RAMDeck runs in offline mode by default — inference calls are restricted to loopback and private-network (RFC1918) addresses unless you explicitly allowlist an external host. Nothing reaches the internet unless you deliberately configure it to.

Q: What is considered proprietary about RAMDeck? Does it use a cloud coordination layer or waitlist?

No cloud layer and no waitlist—RAMDeck operates 100% locally on your home or office network (plus optional encrypted remote access via built-in Tailscale) and is 134% funded on Indiegogo with direct orders open now. Developed by TradeMAV LLC, the proprietary RAMDeck ecosystem combines:

• The RAMDeck Hardware Hub: A dedicated, pre-flashed router appliance (with 5V/3A power supply and Ethernet cable) that hosts your live browser dashboard and OpenAI-compatible API completely offline.
• The Official Zero-Install RAMDeck Node USB ($29 Value Included Free): Our multi-OS thumb drive that activates Windows, macOS, and Linux computers in one double-click with zero installation and lets you physically unplug the USB stick to activate every other computer in your house (watch demo).
• Official Apple App Store & Google Play Mobile Apps: Native iOS/iPadOS and Android apps that pool mobile unified memory and GPU shaders over Wi-Fi via your hub’s LAN IP address.
• Patent-Pending 5-OS Orchestration, 40-Hour Self-Healing, Remote GPU Pooling & Multi-User Knowledge Base: Automatic cross-OS memory coordination, real-world household disconnect recovery (40-hour test), Remote GPU pooling over built-in Tailscale (31 tok/s test), and a private RAM-resident document vault (multi-user RAG demo).

Q: How does RAMDeck compare to free command-line tools, ExLlamaV2, or Petals?

RAMDeck is a complete, plug-and-play hardware and multi-OS ecosystem—not a command-line script. Saying you don’t need RAMDeck because raw developer tools exist is like saying you don’t need an iPhone because microchips, screens, and open-source operating systems can be bought separately on the open market. Could a systems engineer spend a weekend in command-line terminals trying to wire computers together? Maybe—and it would still lack the capabilities that only RAMDeck provides:

• Zero-Install “Unplug & Walk” Node USB (Included Free — $29 Value): Only RAMDeck includes a physical multi-OS thumb drive that lets you double-click once to pool a Mac, Windows, or Linux computer with zero installation, physically unplug the USB stick, and walk it around your house to activate every other computer (watch demo).
• Official Apple App Store (iPhone/iPad) & Google Play Mobile Pooling: Only RAMDeck lets you pool iPhones, iPads, and Android phones over Wi-Fi via official App Store apps (28 tok/s on 30B).
• Dedicated Router Hardware Hub & 40-Hour Self-Healing: Unlike fragile terminal scripts that crash when a laptop closes its lid, or ExLlamaV2 (which only works on a single PC with multiple NVIDIA GPUs), or Petals (which sends your private prompts across strangers’ computers on the public internet), RAMDeck’s Hardware Hub sits on your router and keeps your cluster running 100% privately inside your home or office (read full comparison).

Q: What devices can actually join my cluster?

Windows, macOS (both Apple Silicon and Intel Macs), Linux (Ubuntu, Debian, Fedora), iOS & iPadOS (iPhones and iPads via our official Apple App Store app), and Android (via Google Play). You can connect devices permanently using our 1-click apps or temporarily with zero installation using the included Official RAMDeck Node USB.

Q: How do I get the iPhone / iPad companion app?

The RAMDeck Distributed AI app is available on the Apple App Store (iOS 17+ and iPadOS). Simply download the app, open it while on the same Wi-Fi network as your RAMDeck cluster, and connect via your hub’s LAN IP address to begin contributing high-bandwidth unified RAM and Metal GPU shaders immediately. All data remains 100% private and air-gapped on your local network. You can also read our deep dive: How to Pool iPhone and iPad RAM for Local AI.

Q: How much RAM do I actually need to run a given model?

It depends on the model tier. Your catalog spans Starter (Llama 3 8B needs 6GB pooled RAM, 1 node), Pro (Qwen3.6 27B needs 20GB pooled RAM, 2 nodes), up to Max (a 235B-parameter model needs 128GB pooled RAM across 3+ nodes). The dashboard checks this live against your actual connected devices and tells you exactly which models you currently qualify for.

Q: Is my phone safe to use as a node — won’t it kill the battery or overheat?

Both our iOS and Android companion apps are architected with strict hardware preservation guards. On iOS, the official App Store app was reviewed and approved under Apple Guideline 2.4.2 (Hardware and Battery Protection): it strictly requires the device to be charging and connected to local Wi-Fi, automatically manages unified memory headroom to prevent OS memory pressure, and automatically yields if thermal or battery limits are reached. On Android, the app only contributes while connected to Wi-Fi, charging, and above a 20% battery floor. Phones are designed as auxiliary accelerators or spillover nodes, built to protect your device first.

Q: Are the performance numbers real, or marketing figures?

Real, and we ship the benchmark tool specifically so you don’t have to take our word for it. Every run gets logged with a manifest and per-prompt results, and we’ve publicly flagged our own anomalous readings — one early run showed a suspiciously high 1,065 tok/s that the tool itself caught as a cache-hit artifact, not genuine throughput, and we called it out rather than quietly using it in marketing. You can inspect all verified runs across our test fleet in our empirical report: What Speed Looks Like on a Pile of Old Computers →

Q: Can I plug RAMDeck into tools I already use, like VS Code or my own scripts?

Yes — RAMDeck exposes a standard OpenAI-compatible API (chat completions, streaming, tool/function calling), so anything that already supports a custom base URL — the OpenAI Python SDK, Continue.dev, Cline, or your own code — can point at your cluster with just a URL and API key swap.

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
Scroll to Top