Overview

OpenResearch is a local-first workspace designed to transform your coding agents into powerful research agents. It empowers tools like Claude Code, Codex, OpenCode, and Cursor to conduct in-depth literature reviews, formulate hypotheses, execute experiments, and generate comprehensive research artifacts, all from your local environment.

Key Features

• Agent Versatility: Seamlessly integrate and utilize your preferred coding agents (Claude Code, Codex, OpenCode, Cursor) with customizable harnesses and models per session.
• Parallel Exploration: Assign dedicated, isolated agent sessions to explore multiple research directions concurrently, each within its own Git worktree.
• Reproducible Experiments: Maintain a robust, Git-native experiment tree that meticulously tracks variants and archives each run with an immutable record of its commit.
• Contextual Evidence: Keep all research outputs—logs, diffs, files, results, and artifacts—tightly coupled with the work that generated them.
• Flexible Compute: Choose your compute environment, from local execution and your own infrastructure to managed OpenResearch compute, SSH, Slurm, Kubernetes, Ray, Hugging Face Jobs, Modal, and Tinker.
• Local-First Ownership: All your projects, conversations, experiments, runs, logs, code, and artifacts remain securely on your machine, ensuring local data ownership.
• Autoresearch Capabilities: Enable agents to autonomously navigate the research lifecycle, from proposing ideas and modifying code to launching experiments and analyzing results.

Typical Use Cases

• Accelerated Scientific Discovery: Empower AI agents to autonomously conduct comprehensive literature reviews, identify research gaps, and propose novel hypotheses.
• Rapid Prototyping and Experimentation: Quickly iterate on research ideas by leveraging agents to write code, design experiments, and analyze results in parallel.
• Reproducible Research: Ensure the integrity and reproducibility of your research by meticulously tracking every experiment, code change, and outcome within a Git-based framework.
• Local AI Development: Develop and deploy sophisticated AI research workflows entirely on your local machine, maintaining full control over your data and intellectual property.
• Distributed Research Collaborations: Facilitate seamless research collaboration by enabling agents to run on remote infrastructure while your team interacts with the workspace locally.