HyperAIHyperAI

Command Palette

Search for a command to run...

23 days ago
AI for Science

Open-source AI assistant streamlines research workflow

Lehigh University researchers have developed Dr. Claw, the first open-source artificial intelligence tool designed to manage the complete scientific research workflow. Led by Dr. Lichao Sun, the team created a unified integrated development environment (IDE) that guides users from initial hypothesis generation through to publication-ready results. Unlike existing solutions that require researchers to toggle between multiple specialized applications for literature reviews, coding, or data analysis, Dr. Claw consolidates these functions into a single interface. The system is powered by three leading large language models: Anthropic's Claude Code, Google's Gemini CLI, and OpenAI's Codex. Dr. Claw moves beyond passive content generation to enable autonomous auto-research loops. In this model, the AI actively generates, tests, measures, and learns to improve outcomes, allowing human researchers to focus on high-level strategy and guidance. This approach aims to significantly reduce the time and cost associated with traditional research methods. Addressing concerns regarding data security and accuracy, Dr. Claw processes sensitive information locally or on secure academic servers to ensure proprietary data remains under the user's control. To prevent AI hallucinations, the system avoids making unverified guesses. Instead, it grounds its outputs by cross-referencing peer-reviewed databases and incorporates human-in-the-loop checkpoints to validate critical steps. Dr. Sun emphasized that the goal is to eliminate the inefficiencies of switching tools, a problem he observed frequently among his own students. Remarkably, the software was built by Dr. Sun and five Ph.D. students in just three months, a timeline that would typically require a team of twenty to thirty people working for one to two years. Since its public release in mid-March, the project has gained significant traction on GitHub, approaching 1,000 stars. Early adopters have reported substantial efficiency gains, with one student completing a paper for a top-tier conference in two weeks, a task that previously took two to three months. The rapid development of Dr. Claw has already attracted attention from major institutions. Mayuresh Kothare, Lehigh University's associate dean for research, noted the crowded market of corporate AI tools from companies like Microsoft and Google. He highlighted the achievement of an academic lab developing a competitive open-source tool in such a short timeframe. Dr. Sun's contributions also earned him a spot on the advisory board for the 2026 AI Scientists Conference in Toronto. Looking ahead, Dr. Sun aims to scale the adoption of Dr. Claw globally. He envisions the tool becoming a standard resource supported by university libraries, allowing students to pursue high-risk, long-term projects that were previously unfeasible due to time constraints. By compressing research timelines from years to months, Dr. Claw seeks to fundamentally change the landscape of scientific discovery and accelerate the pace of innovation.

Related Links