Chat with locally installed AI models without sending any code to the cloud.
Let an AI agent read, edit, and run commands on your project files with your approval.
Use a library of ready-made developer prompts for coding, debugging, and documentation tasks.
| riadh35/devai | 0xkinno/neuralvault | 0xlocker/d17-contracts | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | — | TypeScript | Solidity |
| Setup difficulty | moderate | hard | hard |
| Complexity | 3/5 | 4/5 | 5/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires installing Ollama and downloading at least one local model first.
DevAI is a desktop AI assistant for developers that runs entirely on your own Windows computer, with no cloud service involved at any point. It combines a chat interface with local models served by Ollama, an agent that can actually take actions like reading and writing files, running commands, and browsing the web, plus a memory system and a library of ready made prompts. Because everything happens locally, your code, conversations, and data never leave your machine, which the project positions as a fit for sensitive code, offline work, or situations where sending data to a third party service is not allowed. The agent follows a reasoning loop and can call a set of tools such as file editing, command execution, web search, code analysis, and reading PDF or CSV files. It offers three levels of autonomy: read only mode where it takes no actions, a mode where it proposes actions for you to review, and a mode requiring your confirmation before every risky step. Before the agent modifies any file, it saves a backup copy so changes can be undone from a dedicated screen. The project also includes dozens of pre-written prompts organized by category, such as coding, debugging, testing, and documentation, which can be inserted into the chat with one click, and users can save their own custom prompts too. DevAI keeps a memory of each project across sessions, building up facts, conventions, and known pitfalls it learns over time, though any new lesson is only added after the user approves it, never silently. It automatically routes different kinds of tasks to whichever installed Ollama model suits them best, for example a coding focused model for programming questions and a faster model for simple ones. A diagnostics tab shows the health of the backend, Ollama, and other components, along with token usage broken down by day and by model. Installing DevAI means first installing Ollama and pulling at least one model, then downloading either a standard installer or a portable version of DevAI itself, which opens as a local web page in your browser. The tool needs no separate Python installation, is small at around 130 megabytes, and runs on either CPU or GPU, though a GPU speeds things up. It requires Windows 10 or 11 and an internet connection only during the initial Ollama and model download.
A 100% local AI dev assistant for Windows that chats and takes actions on your files using Ollama models, with no cloud service involved.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.