mbro4123

@openclaw Thank you!

@mbro4123 discord.gg/openclaw

@openclaw where's the discord?

We need less updates but more stability. @openclaw

@mytechceoo @albertadevs @getaxal @cisco right now lol

CEO obsessed with token maxxing

@ollama @Kimi_Moonshot We want the local through Ollama

Kimi Kimi 。 Kimi Kimi Kimi Kimi Kimi Kimi ollama run kimi-k2.6:cloud

@ollama Is local coming soon?

Kimi K2.6 raises the bar for open-source models. 🦙 available on Ollama's cloud! Try it with OpenClaw: ollama launch openclaw --model kimi-k2.6:cloud Try it with Hermes Agent: ollama launch hermes --model kimi-k2.6:cloud Try it with Claude Code: ollama launch claude --model kimi-k2.6:cloud more integrations 🧵

@ollama @Thorarnejo Can’t wait!

@Thorarnejo today! doing final checking

Local AI Home Lab Setup Guide - What's Your Stack? The local AI scene is blowing up! Here's what the community is running right now: The Starter Stack: • Ollama (CLI + API) + Open WebUI • Qwen3.5 or Llama 3.3 for general use • RTX 3060+ (8GB VRAM) for 7B-10B models • RTX 4090/PRO 6000 for 70B+ quantized The Pro Stack: • vLLM/SGLang for production workloads • LangGraph + ChromaDB for RAG • Multi-node K3s clusters for scale Power Users: Proxmox VMs + GPU passthrough Beginners: LM Studio (drag-and-drop models) What's your current setup? Drop it below and let's compare notes! ️ My Setup: Hardware: NVIDIA RTX 6000 Blackwell (96GB VRAM) + 192GB System RAM Model: Running qwen3.5:122b (81GB) locally via Ollama Context: Pushed to 200k tokens Software: Linux, Ollama, custom ~/.bashrc tweaks for Flash Attention I haven't ventured into the complex orchestration stacks (vLLM, K3s, etc.) yet. Just kept it simple with Ollama and let the raw VRAM handle the heavy lifting. Curious—how is everyone else managing 70B+ models on smaller GPUs? Any tricks for squeezing out extra context or performance? #LocalAI #Ollama #Blackwell #LLM #HomeLab #AI #NVIDIA #Qwen35 #OpenSourceAI

We're deep in the local AI trenches ourselves—running 120B+ models on Blackwell, managing GPU fleets, building local agent workflows. The power of local is undeniable: no throttling, full privacy, zero API fees. But it requires the right infrastructure to be practical at scale. Would love to test what you're building and share our learnings. The community needs more voices pushing this forward. DM open? Happy to swap notes on stacks, tools, and what's actually working.

Who is running local models on GPUs on OpenClaw? I have started benchmarking different models this week. I am working on improving model selection and switching UX on OpenClaw, i.e. I run /model vllm/gemma-e4b to switch the model in a channel, and then a model controller automatically loads that into memory, gets it ready, or gives an insufficient memory error, if capacity is not enough for that. Like when you are using multiple models in parallel I am going to try llama-swap, LM Studio and Ollama for this next and compare them. There are a ton of variants of models, weight formats and quantizations, which need benchmarking I have been using unquantized original safetensors until now, which already gave me the ability to run ~5 parallel generations in my hardware So if I am going to try LM Studio, I would rather use the bf16 ggml-org/gemma-4-E4B-it-GGUF instead of anything smaller --- because there is no point in nerfing an already smol model if your hardware can run 5 parallel sessions on the unquantized version Will also release vibe reports and benchmarks on all this with @mervenoyann later this week I would like to hear your thoughts if you have already tried these models on OpenClaw

@cybertruck Lol, I actually like it

Heard you wanted a normal-looking truck

@openclaw Let's go!

OpenClaw 2026.4.15 🦞 🤖 Anthropic Opus 4.7 support 🗣️ Gemini TTS in bundled 🧠 Slimmer context + bounded memory reads 🔧 Codex transport self-heal, safer tool/media handling ✨ Pile of update/channel fixes Good boring release. github.com/openclaw/openc…

Hey @Ubiquiti! Loving the UniFi OS 3.x platform 🙌 Quick question: The API docs show full CRUD operations for networks, VLANs, and firewalls, but in practice it seems read-only (POST/PUT/DELETE return 405). Would love to automate network provisioning and management via API – seems like a great feature for power users and enterprise customers! Any timeline for full API write support? 🚀 #UniFi #NetworkAutomation #API #SmartHome

@Cocoanetics Hey, are you (or anyone else) currently using your UniFi skill to create new networks and VLANs? My agent is having trouble with the API keys — it can read but struggles with writes. Getting a lot of CSRF/token issues on UniFi OS 3.x (UDM Pro Max). Any tips or known working approaches? Would really appreciate any guidance. Thanks!

@hexxagon_io Appreciate it @hexxagon_io. Definitely keeping an eye on the resource overhead, but that's the trade-off I'm happy to make for total ownership. No latency, no API limits, and the data never leaves the room. The local loop just hits different. 🦞

Setting up a pure sovereign infrastructure with OpenClaw and UniFi is a solid approach. Make sure to monitor network performance and resource utilization to optimize your setup. At Hexxagon AI, we find tuning configurations for local agents enhances both efficiency and security in production.

Pure sovereign infrastructure. 🛡️ UniFi + OpenClaw local agent. No cloud dependency, no subscription fees, no data leaks. Just raw compute and a clean network. This is the endgame for home AI. 🦞💻 #OpenClaw #UniFi #SovereignAI #LocalLLM #HomeLab

@steipete @maxkolysh My agent kept telling me I needed to set something in the web ui

@maxkolysh docs.openclaw.ai/gateway/secrets

by far the most painful part of using hermes (and openclaw) is secrets management. who is solving this?

@openclaw I usually upgrade, then I run doctor fix. Should I run repair first?

If you have not updated in a while, please don't forget to run: openclaw doctor --repair This keeps your lobster nice and happy and avoids any issues.🦞

OpenClaw 2026.4.11 is out ✨ big polish drop for stability 🛡️ safer provider transport/routing 🤖 more reliable subagents + exec approvals 💬 lots of Slack / WhatsApp / Telegram / Matrix fixes 🌐 browser + mobile cleanup a chunky cleanup pass 😎github.com/openclaw/openc…

@jdrhyne @openclaw Yeah I like the challenge so I upgrade and see what breaks 😂

@mbro4123 @openclaw I am. Just ask your claw to gauge whether the issues you are having will be resolved or not by the new update.

@thsottiaux Is this better than the business tier?

We did it, say hi to the $100 plan! It should be the sweet spot for a ton of you. It comes with a ton of codex usage. And yes we are resetting the limits again too as I mentioned yesterday. Let’s keep building!