Skip to content

DeepCode

Open-source AI coding agent that edits code, runs tools and tests, and resumes long tasks.

Stars
16.8k
Forks
2.2k
Licence
MIT
Language
Python
Last update
4 days ago
Type
AI model / agent

What it is

DeepCode is an open-source coding agent meant for real software projects. You give it a small change or a larger goal, and it works through the project: reading code, editing files, running commands and tests, and checking results. The README says it can continue after an interruption, restart, or model switch. It offers terminal (TUI), desktop, and web clients that share one local service.

  • Ask the agent to make a scoped change to an existing project and review the resulting diff
  • Have it run tests or builds to verify a change it made
  • Run a longer goal that spans many steps and resume it after a restart
  • Use the desktop or web client to watch tool activity and code changes as they happen

Great for

  • Developers who want an open-source agent they can run on their own project
  • People working on multi-step coding tasks that need verification with tests or builds
  • Users who want to inspect diffs and task artifacts rather than accept suggestions blindly
  • Researchers interested in agent harness and multi-agent orchestration ideas tied to the linked paper

Not ideal for

  • Non-programmers who want a no-code tool; the README assumes familiarity with software projects and terminals
  • Anyone who cannot run Python 3.12 or newer, since the README lists that requirement
  • Teams needing documented enterprise support or guarantees; the README does not describe any
Step-by-step guide: how to use DeepCodeInstall, first run, examples, tips and FAQs

How to get it

  1. 1

    See the README

    You'll need: Python 3.12+

    Clone the repository

    git clone https://github.com/HKUDS/DeepCode
  2. 2

    Run it

    • Install DeepCode following the Quick start section of the README
    • Run `deepcode` to open the terminal client
    • Give it a small, specific task on a project you have locally
    • Review the diff and verification output it produces
  3. 3

    Get the most out of it

    • Use the desktop client to see sessions, goals, and code changes visually
    • Use the web client when you want the same shared service in a browser
    • Keep project engineering rules and Skills in the project so the agent can use them for context
    • Reuse Session history to resume interrupted work

Downloads

Strengths

  • Covers terminal, desktop, and web clients over one shared local service
  • README describes verification through tests, builds, diagnostics, and diffs
  • MIT licensed and actively pushed to, with a linked research paper

Trade-offs

  • README is truncated and the install and usage details were not fully visible, so setup specifics are uncertain
  • Describes many features (Skills, Goals, Automations, memory) whose depth is not explained in the visible text
  • Agent behavior depends on models you configure; the README excerpt does not detail supported providers

How it compares

ProjectStarsTypeLevelUpdatedBest for
DeepCode (this)16.8kAI model / agentintermediate4 days agoDevelopers who want an open-source agent they can run on their own project
OpenHands92.5kAppintermediate4 days agoDevelopers who want coding agents running on a server while their laptop is off
GenericAgent14.4kAI model / agentintermediate4 days agoDevelopers who want a small, readable agent codebase they can modify
oh-my-claudecode40.5kDev toolintermediate4 days agoDevelopers already using Claude Code who want multi-agent workflows
deepagents30.9kLibraryintermediate4 days agoDevelopers building AI agents that run multi-step tasks such as research and writing
pydantic-ai21kLibraryintermediate4 days agoPython developers building LLM-powered features that need structured, typed output

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.