Devo CLI
An open-source coding agent in Rust with a persistent Python workspace, child agents and any model provider
- 341 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- OS sandboxing and the permission picker are disabled: the agent runs commands and writes files with the user's full rights
- Project code and context go to the chosen model provider under your key
Install
Manual install
curl -fsSL https://raw.githubusercontent.com/7df-lab/devo/main/install.sh | shFor Linux and macOS. On Windows run the irm install.ps1 | iex command in PowerShell. Run devo in the working folder, and the first launch opens provider setup.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- OS sandboxing and the permission picker are disabled: the agent runs commands and writes files with the user's full rights
- Project code and context go to the chosen model provider under your key
This is third-party code. Review the repository files before installing.
What it does
Devo is a standalone coding agent that carries out multi-step work through a persistent Python workspace and child agents. It runs as a terminal TUI and CLI or as a desktop app, both on top of a local Rust runtime. The ipython tool keeps variables and intermediate results between calls, so the agent can filter large outputs in code and reuse earlier work. The agent delegates independent subtasks to child agents, each with its own session and Python state. The rest is familiar: MCP, skills, permissions and resumable sessions through `devo resume`. You choose the model provider yourself. Endpoints with OpenAI-compatible Chat Completions and Responses or Anthropic Messages connect. The built-in catalog includes DeepSeek, Qwen, Kimi, GLM and MiniMax. The `/model` command switches models, and `devo onboard` repeats the first-run setup.
Who it is for. For developers who want an open agent like Claude Code with any model provider, including local endpoints and private gateways.
Good fit when
- You want a full coding agent for the models in Devo's built-in catalog
- The task consists of steps where it pays to keep intermediate data in Python variables
- The work splits into independent parts for child agents
- Your models live behind a corporate gateway or on a local server
- You want both a terminal and a desktop app on one runtime
Not a fit when
- The agent must be isolated from the rest of your system and files
- You need a stable interface: the project is pre-1.0 and APIs and settings may change
- You have no provider key or local model server: Devo works through your own endpoint
Example request
Parse a week of logs in Python, filter payment service errors and split the analysis across three modules between child agentsLimitations
OS sandboxing and the permission picker are temporarily disabled across the repository: every process gets the full rights of the user, and new sessions open in full-access mode by default. The earlier behavior comes back only by flipping flags in the code.
The project is pre-1.0, so the API and settings may change. On Windows, shell commands need a separate component and a one-time sandbox setup approval. Terminal bundles include Node.js and Python. You need a key for the model provider you choose.
How to disable. Close Devo and delete the binary the installer placed. Provider settings and saved sessions sit in the user profile and can be deleted by hand.
FAQ
How do I install Devo?
On Linux and macOS run `curl -fsSL https://raw.githubusercontent.com/7df-lab/devo/main/install.sh | sh`. On Windows run install.ps1 in PowerShell. The desktop app is downloaded from the releases for your system.
Which models does Devo support?
Any through OpenAI-compatible Chat Completions or Responses and through Anthropic Messages. The catalog includes DeepSeek, Qwen, Kimi, GLM and MiniMax, switched with the `/model` command.
What does the ipython tool do?
It keeps a persistent Python workspace: variables and intermediate results live between calls. The agent filters large outputs in code and does not drag them whole into the context.
How do I resume a previous session?
With `devo resume`. Use `devo upgrade` to update.
Is it safe to run Devo right now?
OS sandboxing is temporarily disabled and processes run with the user's full rights. Run it in a separate folder or a virtual machine and watch what the agent does.
Discussion
No questions or reviews yet.
Related
CLI toolsAn open coding agent for the terminal, browser and desktop that works with models from OpenAI and Claude to Ollama
A terminal coding agent built on Codex, tuned for low-cost open models like Kimi K3, DeepSeek and Qwen
A Go coding agent for open models: it turns a chat into tasks on separate branches, checks and merges the result and runs all projects in one window
Experimental open-source AI software engineer modeled after Devin: plans tasks, researches the web and writes code
A teaching terminal coding agent in Python, built as a "nanoGPT for agents": read it, take it apart and fork it
A Rust CLI that runs development as a team of roles on top of an installed Claude Code, Codex or OpenCode: plan, checks and a final report
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| Devo (this entry) | no | 341 | |
| zerostack: lightweight coding agent in Rust | no | 1,727 | a Rust terminal agent with a sandbox and subagents, with no persistent Python workspace |
| Crush | no | 28,851 | Charm's terminal agent with model switching mid-session |
| Qwen Code | no | 28,798 | the Qwen team's terminal agent |
README in short
The author compares Devo with Claude Code and Codex in a table and gives a quick start for desktop and terminal right away. The warning about the disabled sandbox comes early, before provider setup.