7df-lab/devo

Model-neutral agent desktop/runtime for private, enterprise, and OpenAI-compatible / Anthropic-compatible model API. Tested on DeepSeek, Qwen, Kimi, GLM models. Supports MCP, Skills,and local code search.

What it solves

Devo is designed for complex, multi-step tasks that require more than one-off commands. It solves the problem of LLM agents losing context or starting from a blank state between tool calls by providing a persistent, programmable workspace.

How it works

Devo uses a recursive agent architecture where a parent agent can delegate independent subtasks to child agents. Each agent operates within a persistent Python REPL (the ipython tool), allowing it to keep track of variables, imports, and intermediate results across a session. It can execute bash commands, manage asynchronous work, and and connect to external tools via the Model Context Protocol (MCP) and Agent Skills.

Who it’s for

It is intended for users who need an AI agent capable of handling software development and other complex workflows, accessible via a desktop application or a terminal-based TUI/CLI.

Highlights

  • Recursive Delegation: Parent agents can spawn child agents for focused subtasks, each with its own session state.
  • Persistent Python Workspace: A built-in Python REPL allows for selective data inspection and building upon previous steps.
  • Model Agnostic: Supports OpenAI-compatible and Anthropic APIs, including local model endpoints and private gateways.
  • Extensible: Integrates with Model Context Protocol (MCP) servers and repeatable Agent Skills.
  • Flexible Interfaces: Available as a graphical desktop app and a terminal-native TUI/CLI.

Related

  • Project
    sudoprivacy/sudocode

    A Rust-native CLI coding agent for engineers that provides a high-control, model-agnostic environment for terminal-based AI development.

  • Project
    opendev-to/opendev

    An open-source, Rust-based coding agent that uses a compound AI system to orchestrate parallel agents and mix multiple LLM providers per workflow.

  • Project
    xyTom/coding-tools-mcp

    A model-neutral coding runtime using the Model Context Protocol that gives AI agents secure access to read, patch, and execute commands within a codebase.

  • Project
    vinhnx/VTCode

    VTCode is a terminal-based open-source AI coding agent (Rust) that plans, builds, and reviews code with hosted or local models — featuring an interactive TUI, headless exec mode, sandboxed tool execution, resumable sessions, and a self-correcting "HOVER" loop.

  • Project
    Pinvou/pinvou-agent

    Pinvou Agent is a cross‑platform desktop AI‑assistant that lets you work, design, or code with a conversational model. It provides persistent workspaces, file attachments, a local knowledge base, persona cards, and an extensible plug‑in system (Skills, MCP servers, CLI tools). The UI is built with React + Vite, wrapped by Tauri 2, while the Rust‑based CodeWhale engine handles model calls and tool execution. It runs with any OpenAI‑compatible endpoint or a local vLLM server, and all user data stays under `~/.pinvou3/`. MIT‑licensed and actively maintained.