awesome-deep-research

An open-source platform for creating and deploying autonomous AI agents with pre-built templates.

An open-source autonomous AI agent employing GPT-4 to tackle complex tasks by breaking them into subtasks.

A task management system that creates and executes tasks based on user-defined objectives.

An open-source research community exploring agent scaling laws for collaboration.

A framework for building applications powered by large language models, with components for research workflows.

Framework for building stateful, multi-actor LLM applications extending LangChain.

Presents a framework enhancing LLMs with tools for deep research, outperforming on PhD-level tasks.

Proposes a framework combining MCTS with self-critique for better decision-making in dynamic environments.

Discusses shortcomings in agent benchmarks, proposing cost-controlled evaluations and standardization.

Details on OpenAI's benchmark for browsing agents.

Detailed explanation of Google's approach to deep research with Gemini.

Presents a toolkit leveraging LLMs to help researchers identify, synthesize, and contextualize scientific publications.

Provides specialized AI agents for tasks such as Global News Radar, Scientific Knowledge, and Patent Analysis, particularly useful for research in science and policy domains.

A framework for developing multi-agent conversation systems.

Cooperative AI agents working on research-driven tasks.

A framework for building collaborative AI agents.

A minimal implementation (<500 LOC) of a deep research agent that combines search engines, web scraping, and LLMs to perform iterative research on any topic.

A powerful research assistant that leverages the Motia Framework to perform comprehensive web research on any topic and question.

A step-by-step guide using AutoGPT to create a research assistant that generates reports.

An open-source browser interface powered by Gradio that enables deep research functionality.

A DeepLearningAI course on building research assistants that can reason over multiple documents and answer complex questions, covering routing, tool use, multi-step reasoning with memory, and debugging.

Guide for creating multiple GPT assistants for research using Autogen.

A guide explaining how to automate multi-step searches and summarization using Dify agentic workflow, with iteration nodes that loop through search rounds.

A detailed guide to creating a text analysis agent for research.

Fact Extraction and VERification dataset, for evaluating fact-checking in research.

A dataset for multi-hop question answering across documents, requiring reasoning over multiple sources.

Machine reading comprehension dataset, useful for evaluating information retrieval capabilities.

A benchmark for open-domain question answering, testing agents' ability to find information from the web.

A dataset for question answering over scientific papers, ideal for evaluating research agents in academic contexts.