Devika AI
Autonomous software engineer capable of understanding high-level instructions, breaking them down into steps, conducting research, and writing code to achieve the set goal.
Description
The goal of the product is to create an autonomous agent that can perform software development tasks. The target audience includes developers of all levels who require assistance in creating new features, fixing bugs, or developing projects from scratch.
Capabilities and workflow include the ability to understand complex human requests, decompose them into a sequence of actions, conduct research, and generate code in several programming languages. The agent uses planning and reasoning algorithms, as well as web-surfing capabilities for data collection.
Access and limitations include the use of large language models such as Claude 3, GPT-4, Gemini, Mistral, Groq, and local LLMs via Ollama. API usage requires setting up keys for various models. The project is open source.
The product can be used for automated code generation, bug fixing, project management, creating educational tools, and deploying static websites. To begin work, cloning the repository, installing dependencies, and setting up the environment are required.
