Awesome Deep Research

An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message gateway, it handles different levels of tasks that could take minutes to hours.

No fortress, purely open ground. OpenManus is Coming.

An autonomous agent that conducts deep research on any data using any LLM providers

Tongyi Deep Research, the Leading Open-source Deep Research Agent

"AutoAgent: Fully-Automated and Zero-Code LLM Agent Framework"

~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.

PraisonAI 🦞 — Hire a 24/7 AI Workforce. Stop writing boilerplate and start shipping autonomous self-improving agents that research, plan, code, and execute tasks. Deployed in 5 lines of code with built-in memory, RAG, and support for 100+ LLMs.

Open Source Deep Research Alternative to Reason and Search on Private Data. Written in Python.

No description.

SOTA search powered LLM

DeepResearchAgent is a hierarchical multi-agent system designed not only for deep research tasks but also for general-purpose task solving. The framework leverages a top-level planning agent to coordinate multiple specialized lower-level agents, enabling automated task decomposition and efficient execution across diverse and complex domains.

Humanity's Last Exam

"Your Fully-Automated Personal AI Assistant"

[NeurIPS 2025] 🌐 WebThinker: Empowering Large Reasoning Models with Deep Research Capability

ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning & ReCall: Learning to Reason with Tool Call for LLMs via Reinforcement Learning

ZeroSearch: Incentivize the Search Capability of LLMs without Searching

🔍 Search-o1: Agentic Search-Enhanced Large Reasoning Models [EMNLP 2025]

Salesforce Enterprise Deep Research

An Open-Source AI Writing Project.

[EMNLP'25] s3 - ⚡ Efficient & Effective Search Agent Training via RL for RAG (RLVR for Search with Minimal Data)

CORAL is a robust, lightweight infrastructure for multi-agent autonomous self-evolution, built for autoresearch. Works with Claude Code, Codex, Cursor, OpenCode, Kiro, and more.

DeepResearch Bench: A Comprehensive Benchmark for Deep Research Agents

Scaling Deep Research via Reinforcement Learning in Real-world Environments.

R1-searcher: Incentivizing the Search Capability in LLMs via Reinforcement Learning

Dr. Zero Self-Evolving Search Agents without Training Data

[ACL-2026] MMSearch-R1 is an end-to-end RL framework that enables LMMs to perform on-demand, multi-turn search with real-world multimodal search tools.

recursive rag with r1 reasoning

Deep Research

A Deep Research agent from scratch

MrlX: A Multi-Agent Reinforcement Learning Framework

No description.

Repo for "MaskSearch: A Universal Pre-Training Framework to Enhance Agentic Search Capability"

No description.

Official repository for RAG-Gym

The official repo of "WebExplorer: Explore and Evolve for Training Long-Horizon Web Agents"

SimpleDeepSearcher: Deep Information Seeking via Web-Powered Reasoning Trajectory Synthesis

[NeurIPS'25 D&B] Mind2Web-2 Benchmark: Evaluating Agentic Search with Agent-as-a-Judge

[EMNLP 2025] WebAgent-R1: Training Web Agents via End-to-End Multi-Turn Reinforcement Learning

R1-Searcher++: Incentivizing the Dynamic Knowledge Acquisition of LLMs via Reinforcement Learning

A High-Efficiency System of Large Language Model Based Search Agents

IKEA: Reinforced Internal-External Knowledge Synergistic Reasoning for Efficient Adaptive Search Agent

Source code of paper: Process vs. Outcome Reward: Which is Better for Agentic RAG Reinforcement Learning

No description.

[AAAI 2026] AutoTool: Efficient Tool Selection for Large Language Model Agents

Official Implementation of "O-Researcher: An Open Ended Deep Research Model via Multi-Agent Distillation and Agentic RL"

HiPRAG (Hierarchical Process Rewards for Efficient Agentic Retrieval Augmented Generation) is a reinforcement learning method designed for training reasoning-and-searching interleaved LLMs with improved efficiency and reduced oversearching as well as undersearching behavior.

WebSeer: Training Deeper Search Agents through Reinforcement Learning with Self-Reflection

ToolRM: Towards Agentic Tool-Use Reward Modeling

Learn to resolve ambiguity for Question Answering through RL

No description.