What problem does it solve?
Autoresearch eliminates the manual overhead of managing an AI research project by coordinating literature intake, hypothesis formulation, iterative experiments, and final synthesis into publishable outputs.
Core Features & Use Cases
- Two-loop research orchestration: Runs an inner loop for rapid hypothesis experiments and an outer loop for reflective synthesis that updates the narrative and research direction.
- Autonomous routing to domain skills: Delegates domain execution (training/evaluation/data/analysis) to specialized skills while the orchestrator manages state, protocols, and bookkeeping.
- Continuous research operation: Maintains a wall-clock continuity loop (Claude Code /loop and OpenClaw heartbeat) so research keeps progressing across time ticks.
- Research artifacts and human-visible progress: Creates a structured workspace with state, logs, findings, per-hypothesis experiment folders, and progress presentations (HTML/PDF).
Quick Start
Use autoresearch to start a new research project from a question by first creating the workspace and then letting it run the mandatory continuity loop before beginning experiments.