Agentlas OS
Agentlas OS is an open-source platform for building, borrowing, and owning AI agents. It turns plain-language requests into runnable agent or team packages that are portable across hosts and can be stored in a private Agent Cloud. It is for developers and teams who want to create reusable agents not tied to a single model or workspace.
✨ Key features
- Build agents from plain-language requests with roles, tools, and permissions.
- Borrow specialist agents from the public Agentlas Hub.
- Own agents in a private, owner-scoped Agent Cloud.
- Portable packages run on any supported host with your chosen LLM.
- Verification gates ensure packages meet required contracts.
- Install via paste-to-install for LLM hosts or shell command.
🎯 Use cases
- Create a reusable agent from a working session method.
- Orchestrate a team of local and borrowed agents for complex tasks.
- Move an agent between computers or model providers.
- Share agents publicly via the Hub while keeping source private.
- Run agents locally with your own model and credentials.
📦 Installation
🧰 Requirements: Runs on macOS (Apple silicon and Intel), Windows (via Git Bash), and Linux. Requires curl and bash for installation; no API keys or accounts stated.
Paste to Install (inside an LLM host)
Paste this into the LLM you are using (e.g., Claude Code, Codex, Gemini CLI, Cursor):
I want to install Agentlas OS / Hephaestus from this open-source GitHub repo:
https://github.com/agentlas-ai/Agentlas-OS
Before running anything, fetch and read the installer script so you know what
it does:
https://raw.githubusercontent.com/agentlas-ai/Agentlas-OS/main/scripts/install-all-runtimes.sh
It downloads a release tarball from this same repo's GitHub Releases and
writes files only under ~/.agentlas, ~/.local/bin, and this host's own plugin/
command-adapter directories (e.g. ~/.claude for Claude Code) — nothing outside
those paths. If that matches what you read, run it:
curl -fsSL https://raw.githubusercontent.com/agentlas-ai/Agentlas-OS/main/scripts/install-all-runtimes.sh | HEPHAESTUS_INSTALL_GLOBAL_ROUTER=1 bash
HEPHAESTUS_INSTALL_GLOBAL_ROUTER=1 additionally writes a routing block into
this host's global instructions file (e.g. ~/.claude/CLAUDE.md) that lets
substantial tasks be staffed from Agentlas' agent network; it is not secret
and you can read/quote it back to me. Drop that env var if you'd rather skip
it for now — you can a
🚀 Usage
Build an owned agent
Describe the work in plain language. Agentlas classifies the request, runs the interview and research gate, generates the package, verifies it, then asks whether to keep it only on this computer or save it privately in Agent Cloud.
Turn a working session into an owned agent
In an interactive host, use the canonical session route:
hep-build session
Agentlas asks for the destination first; if you do not name another folder, it creates a new safe-slug child below AGENTLAS_AGENT_HOME or, by default, ~/.agentlas/agentlas-agent. Existing children are never overwritten. It then shows a generalized session report for your review, turns the approved method into a standalone agent prompt and package, registers it locally, and runs the normal verification gates.
⚠️ Good to know
The README notes that Agentlas is not a claim of regulated financial or legal trust services, and that credentials, local files, and machine-specific permissions do not travel with the package—you configure those separately on each computer.
❓ FAQ
How do I install Agentlas OS?
You can paste the provided install text into an LLM host like Claude Code, or run the curl command directly in a terminal (Git Bash on Windows, Terminal on macOS). The installer downloads a release tarball and writes files only under ~/.agentlas, ~/.local/bin, and plugin directories.
Can I use my own LLM with Agentlas OS?
Yes, you use the model account or API key you choose. The current host performs the work under its local files, tools, credentials, permissions, and verification rules.
What is the Agent Cloud?
Agent Cloud is a private, owner-scoped storage for agents you create. It stores and retrieves your package, but it is not a hosted LLM; your chosen model and current host execute the work.
How do I borrow agents from the Hub?
Use /agentlas hub (Hub only) or /agentlas network (Local + Cloud + Hub) to find public Hub specialists and bring the selected runtime bundle into your current Agentlas host. The publisher's private source work is not copied into your workspace.
📊 Repository
🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.