OpenManus setup, and the open-source system a team moves to
OpenManus setup is a documented Python install you run on your own machine, and Kortix is the open-source AI Operating System you move to when the work has to belong to a team. Run the install, then see where a framework stops and what a repo the whole team owns picks up.
Two ways into the same repo
Real Kortix text from the project README.
$ kortix init$ kortix ship$ kortix sessions new --prompt "Summarize this week's commits and open a change request"$ kortix self-host start$ kortix cr ls # review what an agent proposesWhat OpenManus setup is
OpenManus is an open-source Python framework for building general AI agents. Its project site uses that description, and its repository holds the setup: clone the repo, build a Python 3.12 environment, install the requirements, write a config/config.toml with a model key, and run it.
The OpenManus repository documents two install paths. The conda path creates an environment named open_manus on Python 3.12, clones the repo, and installs with pip install -r requirements.txt. The uv path installs uv, creates a virtual environment with uv venv --python 3.12, and installs with uv pip install -r requirements.txt. The project recommends the uv path for faster installation and better dependency management.
OpenManus configuration is one file. Copy config/config.example.toml to config/config.toml and set the [llm] block with the model, base_url, and api_key you want to reach. The general agent starts with python main.py; the repository also documents python run_mcp.py for its MCP tool version and python run_flow.py for the multi-agent path it labels unstable. Browser automation runs as a default MCP server through uvx browser-use --cli-mcp.
Where the setup stops
Everything OpenManus setup touches stays on the machine you run it from: the Python environment, the config file, the model key, and every artifact the agent produces. OpenManus is a framework for building agents, so it does not ship a shared repo, a review step before a change takes effect, a schedule or webhook that starts work when nobody is present, or a hosted path for a team.
The missing piece shows as soon as a second person joins. A teammate cannot read the agent's memory, approve its work, or run the same configuration on another machine without repeating the install. OpenManus teaches one person how the loop works, and the loop stays with the one person who installed it.
What a team owns instead
Kortix is the open-source AI Operating System. Your agents, skills, company memory, and connectors live in one git repo you own, and each session gets its own isolated Linux sandbox and branch. A session clones the repo and works on a new branch, the agent opens a change request against the default branch, and a person merges it. The shared repo, the branch, the review, the memory, and the connectors are the parts a framework leaves to you.
You run Kortix on Kortix Cloud, in your VPC, or on your own on-prem network, and self-host is free. Any model provider with your own keys.
If you are still weighing the options, the OpenManus alternatives guide on this site sets them side by side, the self-hosting guide covers running the whole system on your own infrastructure, and the home page starts from the decision.
The parts a team gets as files
Each layer maps to real Kortix configuration.
- Repo
The company is a git repo
A project is a git repo with
kortix.yamlat its root, and agents, skills, memory, tools, triggers and connectors are files in it.project project (git repo + kortix.yaml) └─ session ──> cloud computer: an isolated sandbox on a branch named after the session └─ the OpenCode agent works └─ change request ──> you review & merge ──> mainSource: https://kortix.com/docs - Agents
Agents are markdown files
An agent is
agents/<name>.mdplus anagents.<name>block inkortix.yaml, the one for behavior and the one for governance.kortix.yaml # kortix.yamlagents: support-triage: file: agents/support-triage.md # what it does, in plain markdown connectors: all # what it may reachSource: https://github.com/kortix-ai/suna - Triggers
A trigger starts a session
A trigger starts a session on a cron schedule or a signed webhook, with nobody present.
kortix.yaml # kortix.yamltriggers: - slug: daily-digest type: cron cron: '0 0 9 * * 1-5' prompt: Summarize yesterday's support tickets and open a change request with the digest.Source: https://github.com/kortix-ai/suna
OpenManus setup questions
Short answers, each from OpenManus's own repository or project site.
No. OpenManus is a public repository you clone and run yourself, and its project describes building agents without an invite code. There is no account to request first.
It is a framework. Its own site calls it an open-source framework for building general AI agents, so you get the building blocks and a running agent rather than a hosted service with accounts and a dashboard.
The example config ships with gpt-4o at https://api.openai.com/v1. You copy config/config.example.toml to config/config.toml and replace the placeholder api_key with your own.
Kortix, the open-source AI Operating System: your agents, skills, company memory, and connectors in one git repo you own, with each session on its own isolated Linux sandbox and branch.
Give the work a repo your team owns
Kortix is the open-source AI Operating System: agents, skills, memory and connectors in one git repo.
Open source · Any model, your keys · Self-host, VPC, or on-prem