Skip to content

Cette section n'existe pas encore en français — le contenu ci-dessous est en anglais. La traduction arrive.

🤖 Agents IA

Les agents IA vont au-delà du chatbot. Ils peuvent exécuter des tâches sur votre système, écrire du code, lancer des commandes et apprendre de leurs réussites comme de leurs erreurs. Cette catégorie suit les frameworks d'agents auto-hébergés et les outils pilotés par des agents.

3 resources — sorted by rating

🤖 Agents IA
Tool
Free

Nous Hermes Agent

An open-source, autonomous AI agent framework by Nous Research with a persistent memory loop that creates and refines its own reusable skills over time.

Best for: Developers seeking a model-agnostic agent framework that learns from experience and builds a persistent memory of user preferences.

Fabian's take — "Typical agents treat every task as a blank slate, which gets annoying fast. Hermes has a built-in learning loop, so it actually builds and saves reusable skills as it works. If it figures out how to refactor a file or parse a document once, it remembers that pattern for the next task. It's model-agnostic too, so you aren't locked into a single provider — for a free, self-hosted framework, that pairing is rare."

Free (Open Source) + your own LLM API costs Read review
🤖 Agents IA
Tool
Free

OpenClaw

A self-hosted, open-source autonomous agent framework that runs locally, connects to your files and systems, and integrates with messaging platforms like WhatsApp and Discord.

Best for: Developers and power users who want an autonomous, self-hosted AI agent with full local file and command access, and chat integration

Fabian's take — "OpenClaw is a fantastic open-source framework for building self-hosted autonomous agents. I love that it connects directly to messaging apps like WhatsApp. You can literally ask it to run updates on your website or search your local documents via chat. But treat it with respect — giving an autonomous agent access to execute terminal commands on your system is like letting a teenager drive your sports car: make sure you've set up sandboxes first."

Free (Open Source) + your own LLM API costs Read review
🤖 Agents IA
Tool
Freemium

Perplexity Computer

Perplexity's cloud-based agentic environment that executes multi-step workflows, writes and runs code, and operates tools autonomously — making vibe coding accessible without local terminal setup.

Best for: Non-developers who want an autonomous agentic workspace to run scripts, gather data, and build prototypes without managing a terminal or local files

Fabian's take — "Perplexity Computer is basically a cloud-based agentic sandbox. If Claude Code is the professional workbench, this is the user-friendly entry point for non-techies. It runs entirely in the cloud, so you don't need to touch a terminal, deal with Node versions, or worry about security sandbox configurations. You just give it a task — like pulling some API data, organizing a dataset, or building a prototype — and let it run. It's incredibly fun to watch it think and figure out its own mistakes."

Included with Perplexity Pro ($20/month) / Enterprise plans Read review

Restez au courant

Ne manquez pas la suite

Je sélectionne les meilleurs outils IA pour les professionnels. Inscrivez-vous à la liste et je vous écrirai quand j'aurai quelque chose qui vaut la peine d'être partagé.La newsletter est en anglais.