Devika CLI
Experimental open-source AI software engineer modeled after Devin: plans tasks, researches the web and writes code
- 19,555 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- The agent writes and runs code and uses a browser
- Stores keys for many services, including Netlify for deploys
Install
Manual install
git clone https://github.com/stitionai/devika.git
cd devika
uv venv
source .venv/bin/activate
uv pip install -r requirements.txt
playwright install --with-deps
python devika.pyThe frontend runs separately: cd ui/, bun install, bun run start
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The agent writes and runs code and uses a browser
- Stores keys for many services, including Netlify for deploys
This is third-party code. Review the repository files before installing.
What it does
Devika takes a high-level task, breaks it into steps, researches the web and writes code. It works through a web UI with projects, chat and agent state visualization. It supports Claude 3, GPT-4, Gemini, Mistral, Groq and local models via Ollama. The authors warn the project is at an early stage and many features are unimplemented or broken.
Who it is for. Enthusiasts and researchers studying autonomous agent architecture.
Good fit when
- You want to study how an autonomous agent with planning and web search is built
- You want to experiment with local models via Ollama
- You are ready to improve the project yourself
Not a fit when
- You need a stable agent for real work
- You cannot set up Python, Node.js, bun and Playwright
Example request
Build a simple Flask to-do web app with SQLite storageLimitations
The project is experimental and many features do not work. Requires Python 3.10-3.11, Node.js 18+ and bun. Web search needs Bing or Google keys. The README points to Opcode as the second iteration.
How to disable. Stop the devika.py server and the frontend, then remove the project folder.
FAQ
Where do I set API keys?
On the settings page in the web UI; values are stored in config.toml.
Which model works best?
The README recommends the Claude 3 family.
How is Devika different from terminal coding agents?
It is an experimental open agent inspired by Devin with a web UI: projects, chat and agent state visualization. It splits a task into steps, searches the web and writes code.
How do I download and run Devika?
Clone the repository with git clone https://github.com/stitionai/devika.git, create an environment with uv venv and install dependencies per the README. API keys are set in the settings. Risk is high and the project is experimental.
Discussion
No questions or reviews yet.
Related
CLI toolsAn open-source coding agent in Rust with a persistent Python workspace, child agents and any model provider
An open coding agent for the terminal, browser and desktop that works with models from OpenAI and Claude to Ollama
A local-first AI agent running on your own machine via llama.cpp that drives a browser, edits files, runs approved commands and remembers context across session
A CLI to run several coding agents: one config, a session console, local memory and a set of engineering skills
A desktop app where Claude Code, Codex, Gemini and other agents work as a team of specialists in a shared project
A multi-agent harness that verifies results by artifacts: stop hooks, an independent judge and an event log for Claude Code, Codex, Cursor and more
README in short
Devika is meant as an open-source alternative to Cognition AI's Devin. The agent understands natural-language tasks, plans, researches and writes code in several languages. The README covers installation with uv and bun, running the web UI and configuring keys in config.toml. It states upfront that the project is experimental and that the next iteration is called Opcode.