AI-Researcher is a fully automated system designed to streamline the entire scientific research lifecycle, from literature review and idea generation to algorithm implementation, validation, and manuscript creation. It targets researchers and power users seeking to accelerate innovation by leveraging advanced AI agents for end-to-end research automation.

How It Works

The system operates in three phases: Literature Review & Idea Generation, Algorithm Design, Implementation & Validation, and Paper Writing. It utilizes a "Resource Collector" to gather materials from academic databases and code platforms, a "Resource Filter" to select high-quality resources, and an "Idea Generator" to formulate novel research directions. The core research pipeline involves iterative design, implementation, validation, and refinement of algorithms. Finally, a "Writer Agent" automatically generates academic papers based on the research outcomes.

Quick Start & Requirements

• Installation: git clone https://github.com/HKUDS/AI-Researcher.git && cd AI-Researcher && pip install -e .
• Docker: Pull tjbtech1/paperagent:latest for the agent-interactive environment.
• API Keys: Required for LLMs (OpenAI, Deepseek, Anthropic, Gemini, HuggingFace, Groq, XAI) via .env.template.
• Hardware: Multiple GPUs are recommended for running the research agent (e.g., GPUS='"device=0,1"').
• Documentation: Comprehensive documentation is "on its way."

Highlighted Details

• Supports two input levels: detailed idea description or reference-based ideation.
• Features a comprehensive, open-source benchmark suite covering CV, NLP, Data Mining, and IR.
• Integrates with multiple LLM providers (Claude, OpenAI, Deepseek, etc.).
• Automates the generation of full-length academic papers.

Maintenance & Community

• Community channels include Slack and Discord.
• GitHub Issues can be used for contributions and discussions.

Licensing & Compatibility

• The repository does not explicitly state a license in the README.

Limitations & Caveats

• Comprehensive documentation is still under development.
• The specific LLM models used in the example commands (claude-3-5-sonnet-20241022, claude-3-5-haiku-20241022) may require specific API access or subscriptions.