Run Private AI agents 24/7 to increase your daily productivity. Earn rewards running your web3 node at home. Zero hassle, it just works https://nitter.cf/t.co/PXEiFrEo9S
Cryptovalley-Zug (Switzerland)
Joined February 2018
- Tweets3.9K
- Following1.3K
- Followers16.1K
- Likes6.6K
Qwen 3.8 27B is the model that packs the highest intelligence per size
It's fast, smart, and very good with agentic tasks
It can handle pretty much everything you throw at him
And it's also Free🤑 for a limited time in Nexus right now
⚡️Go squeeze every prompt from it! 🥵
If you are not running your own agent you are behind
"But it's too complicated"
In 1:47 seconds you can have it running on your Dappnode
Get your agent, fully private & self-sovereign
Comment 👇which thing you'd die to automate in your life and we'll make a video of how
First, thank you also to all the matching fund providers. Matching came from @TheDAOfund @quantstamp @wintermute_t @osec_io @Certora @sigp_io @CertiK @chain_security. Security teams funding security 🤝
Thank you❤️
TheDAO Fund's Ethereum Security round on Giveth sent Dappnode 5.6343 ETH: 1.2227 from the community, 4.4116 matched.
Here's how that money becomes Ethereum security: easier home staking → more solo validators → a more decentralized, safer Ethereum.
Testing on Dappnodes
Today, we're introducing Mach-1 Additive, a 35 billion parameter model that can inference without ever multiplying by a weight. At 1.7 bits per weight, Mach-1 recovers 95% of the performance of the original full precision model, Qwen 3.6 35b, across 12 agentic and reasoning benchmarks, while being 10x smaller.
At 7GB, Mach-1 comfortably fits on consumer laptops with speeds of up to 120 tokens per second, making local inference not just feasible but useful.
Unlike algorithms like BitNet, our approach requires minimal retraining, under 15 GPU hours, making it scalable to massive LLMs. Over the coming weeks, we will be announcing and serving models of up to 3 trillion parameters compressed using our algorithm. For now, you can visit our website to play with Mach-1 directly in your browser, or download our desktop app.
We couldn't be more excited to launch Mach-1. We're looking forward to an energy efficient future for AI, powered by scaled intelligence density.
Your Nexus API key works beyond Nexus Chat: point any
OpenAI-compatible client at `nexus-api.dappnode.com/v1` and you're routing through Nexus.
Works in OpenClaw + Hermes Agent (Dappnode
packages) and OpenCode (open-source dev tool). Plus anything else that takes a base URL + key.
Power users: DAppNode also exposes a `POST /mcp` endpoint, so any MCP client (Claude Desktop, Cursor, custom agents) can talk to your Dappnode directly.
Read-only by default; mutating tools opt-in.
→ Debug: read logs, find the cause
→ Install: search and install packages
→ Configure: env vars, ports, staker settings
→ Lifecycle: start, stop, restart
→ Confirm: every action shows a card before it runs
Nexus Chat isn't a chatbot bolted on.
It runs against your node. It sees your installed packages, your logs, your system info. It's grounded in DAppNode's own docs.
What it can do:
DAppNode 0.3.23 is out
It ships Nexus Chat — a built-in assistant that lives inside the UI.
Reads your logs. Debugs your stack. Installs packages. Asks before it changes anything.
Your node, now with a copilot
So incredibly thankful for your @OctantApp support!
Still 9h for you to assign rewards or donate ETH directly to get HEAVILY matched by Octant!
Help us remain Free and Open Source Forever!
The @OctantApp community helped keep us alive through a rough stretch.
This round is your chance to do it again — and the matching pool is bigger.
Quadratic funding rewards breadth: more unique donors = bigger match. Bring a friend. Bring five.
Octant's QF round closes in 1 day, 8 hours.
200 ETH matching pool. 25 projects. Dappnode is one of them.
Help keep Dappnode Free & Open Source!
Every unique donor gets matched. Even $5 counts.
Donate 👇 (link below)
you don't have to use less AI. you have to use the right one:
nexus.dappnode.com
you don't have to choose between selling your free will and falling behind
we built Nexus so you can use frontier models without logging, training, or profiling.
OpenAI-compatible. swap one URL. ship today.
