Agence ≠ Différente

Dernière réalisation de l'agence en collaboration avec le studio Perméable - refonte du site du cabinet d'architectes AUP aup-architectes.fr

@ggerganov Hello how much RAM ?

Let me demonstrate the true power of llama.cpp: - Running on Mac Studio M2 Ultra (3 years old) - Gemma 4 26B A4B Q8_0 (full quality) - Built-in WebUI (ships with llama.cpp) - MCP support out of the box (web-search, HF, github, etc.) - Prompt speculative decoding The result: 300t/s (realtime video)

Ça ne s’arrête plus

Claude c'est vraiment une dinguerie pour le SEO je viens de trouver comment Claude peut gérer lui-même ton SEO sans te casser la tête et même quand tu dors et c'est parfait si tu vibe codes, peu importe comment J'ai tout mis dans un NOTION détaillé ça t'interesse ? RT+Commente ''Claude SEO'' et je te l'envoie en DM

@TontonJars Claude SEO

@gitlawb ça va pas rester en ligne lontemps - le département légal va vous contacter comme ils l'ont fait pour openclaw (clawdbot)

Set 3 env vars and you're running Claude Code's full tool system, file editing, search, bash, git — powered by whatever model you want.

We forked the leaked Claude Code source and made it work with ANY LLM: GPT, DeepSeek, Gemini, Llama, MiniMax. Open source. The name is OpenCode

@Zoeillle J'ai un 32GB c'est bien mais j'imagine qu'à 64GB tu feras tourner de meilleur modèle. Notamment via Ollama (j'ai un M4, j'ai lu que le M5 était mieux, ce qui est en soit logique)

C'est décidé. En Juin je m'achète un Mac Mini 32GB et je monte un LLM de code dessus. En perso j'aimerais leave Claude éventuellement car le coût à l'usage est devenu beaucoup trop cher. Ce sera uniquement en perso J'hésite juste entre 32GB de RAM ou 64 (prix x2)

Great

@melvynx Openclaw will be the Netscape of personal agents

Went to an "OpenClaw meetup" in SF; literally no one understood this tool, even the people that went to the meetup about it. This is pretty insane. I think we are like 1,000 who understand the power of the tool on earth

@TfTHacker @mreflow a plugin that allows you to send query to Claude Code (like the gemini skills or codex one) github.com/bettep-dev/ope… With the last version, that should not be necessary anymore

@agentdifferent @mreflow What is a Claude bridge?

For those using OpenClaw at a high level, what’s your favorite default model? I was using Nemotron 3-super locally on my Spark but it hits context limits too quickly. I’m mostly using Sonnet-4.6 now but API costs rack up fast. I love my claw but honestly haven’t optimized models and model switching as much as I should. I like the bigger context windows when building out new skills and automations so it doesn’t forget what we’re building but that’s also when API costs soar. I like my local models because I have a Spark for that reason but context windows aren’t great on the local models I’ve tried… Looking for advice from some more experienced OpenClawers.

@ItsAditya_xyz @openclaw @steipete @vincent_koc same same

@openclaw hey @steipete @vincent_koc was the release built without the UI artificats?

OpenClaw 2026.3.22 🦞 🏪 ClawHub plugin marketplace 🤖 MiniMax M2.7, GPT-5.4-mini/nano + per-agent reasoning 💬 /btw side questions 🏖️ OpenShell + SSH sandboxes 🌐 Exa, Tavily, Firecrawl search This release is so big it needs its own table of contents. github.com/openclaw/openc…

@s_mohinii For those interested I've created a boilerplate and a script to easily create a project based on this architecture github.com/nottavi/cc-boi…

Building a "Claude Code" project isn't just about dumping files into a folder—it’s about architecting a repository that an AI can actually navigate without getting lost. ​If your CLAUDE.md is becoming a knowledge dump, you’re doing it wrong. Here is the blueprint for a high-performance AI-native repo: ​📂 The Modular Blueprint ​CLAUDE.md (The Brain): Not a junk drawer. Keep it lean with the project’s Purpose (Why), Repo Map (What), and Working Rules (How). ​.claude/skills/: Move repeated instructions here. Think of these as "playbooks" for code reviews, refactoring, and releases to ensure consistency. ​.claude/hooks/: Models forget; hooks don't. Use these for deterministic automation like running formatters or blocking unsafe directory edits. ​docs/: Use for progressive disclosure. Keep architecture decisions and runbooks separate so the AI only pulls what it needs.

@Voxyz_ai Can I just say defaultProfile=« Profile 2 » ?

openclaw 3.13 added live chrome attach. before this, my agents couldn't see what i see. dashboards behind login walls, authenticated sessions, real browser state. i had to screenshot and paste. now: # openclaw.json { "browser": { "defaultProfile": "user" } } # chrome://inspect/#remote-debugging → enable # or launch with: chrome --remote-debugging-port=9222 enable remote debugging in chrome, and the agent connects to your actual browser. logged in. real cookies. no extension needed. tested locally. told scout to check my vercel dashboard for traffic data. it opened the tab i already had logged in, read the numbers, and wrote them into the weekly report. no headless browser. no separate login. no screenshots. it just sees what i see. the gap between "my agent can browse the web" and "my agent can browse MY web" was always authentication. this closes it. ps: requires chrome running on your local machine. vps users need to connect your machine as a node.

@0xplit @pascal_ccbb 41 agents 🤯🤯 ça doit être galère à administrer En gros tu as créé un agent pour chaque « tâches » ?

@pascal_ccbb Pas beaucoup de skill, 95% il se build tout seul. 41 agents, en parti pour recap auto le matin, monitoring uptime, indexation Google, rédaction/publication/maillage, nettoyage SEO (suppressions d'articles + 301), synchronisation catalogue de vente de liens

Quels sont vos skill préférés sur OPENCLAW et vos utilisations quotidienne ? J'aimerais voir un peu ce qu'il est possible de faire (coté seo, gestion serveur etc...)

@steipete Cool 👍 What about bird, you have removed the repository, did X asked you to shut it down ?

🧭 Shipped gogcli 0.12.0: Google in your terminal, now with Workspace Admin, ADC/access-token auth, Docs tab editing + Markdown/HTML export, huge Sheets upgrade, calendar aliases/subscribe, forms watches and slides templates. brew install gogcli github.com/steipete/gogcl…

@steipete @openclaw All models are struggling with context / memory I use Kimi 2.5 - it is very good but slow - it is confirmed in this benchmark

Interesting benchmark on which model is best for @openclaw pinchbench.com

@Eljaboom Where do we find the original markdown file ?

This is INSANE. Boris Cherny, the head of Claude Code just revealed his exact workflow. This is a goldmine, must repost & bookmark:

@RG22ESP @XenxFN @lixxfn11 @vexifnr Go go

@Loran750 Cela m’a fait écrire un bot telegram pour alimenter mon Shaarli

j'ai éclaté de rire. C'est exactement ça. La plupart des gens qui utilisent OpenClaw vont l'abandonner la semaine prochaine quand ils vont s'apercevoir que ça les fait ch**** au final car ils n'en font rien d'intéressant. Les autres ? Oui, c'est très utile. Mais pas pour tout le monde.

Gemini 3.1 Pro is here. We’ve significantly improved the model’s overall intelligence so it can solve tougher problems. 🧵