THOR

Settings

Appearance

Research Configuration

Choose your research mode and, when available, synthesizer in the research input.

Go to research controls

Developer Diagnostics

PRODUCTION MODE

Research Planners

Planner 1 — Gemini 3.5 Flash
google/gemini-3.5-flash

Planner 2 — Llama 4 Scout
meta-llama/llama-4-scout

Planner 3 — Qwen3.5-27B
qwen/qwen3.5-27b

Retrieval

Semantic Scholar: Enabled
API key: Not configured (public access)

Consensus: Enabled

Synthesizer

Deepseek V4 Flash
deepseek/deepseek-v4-flash

Credential readiness

OpenRouter: Configured

Consensus: Configured

Manual evaluation cases and results

Saved-run diagnostics

Select a saved chat to inspect its run diagnostics.

About THOR

THOR
Thoughtful Higher-Order Research Agent

THOR is a multi-agent academic research assistant designed to help researchers plan searches, retrieve scholarly evidence, compare findings, identify research gaps, and synthesize evidence into traceable research insights.

Pinned chats are saved in this browser.

THOR

Thoughtful Higher-Order Research Agent

Ask one question. Get evidence-grounded research from multiple AI agents and academic databases.

THOR combines multi-agent AI planning with academic evidence from Consensus and Semantic Scholar to help you discover gaps, compare findings, and develop stronger research questions.

Start Research

Find a research gap

Identify underexplored areas in AI-assisted health professions education.

Develop a thesis topic

Generate feasible Master's-level research questions from current literature.

Literature Review

Access and summarize relevant academic papers.

Multi-agent research planning
Academic evidence retrieval
Semantic Scholar
Consensus
Traceable citations
Evidence-aware synthesis

A starting point for critical inquiry. Review the sources and apply your own research judgment.

Three AI perspectives + academic evidence · Check sources before relying on them