ИИ-агенты
local-deep-research
Deep research that stays on your machine — local or cloud LLMs, 10+ search engines, encrypted and private.
★ 9,2 тыс. звёздMIT
Смотреть руководство →Tongyi Deep Research — проект с открытым исходным кодом из категории «ИИ-агенты» от Alibaba-NLP (Alibaba-NLP/DeepResearch на GitHub). У него 20 022 звезды на GitHub и 1 532 форка, он распространяется по лицензии Apache-2.0 и последний раз обновлялся на GitHub 27 февраля 2026.
Tongyi Deep Research - Alibaba's open-source deep research agent and 30B model
Collect the research questions you want to evaluate, in the JSONL format the repo expects.
Clone the repo and download the Tongyi-DeepResearch model weights from the Model Download section.
Create the conda environment, install requirements, and copy the example environment file.
Configure the inference script (model path, dataset, output) and run it.
A ReAct-style agent that plans, searches and synthesizes long-form research answers.
Tongyi-DeepResearch-30B-A3B weights are downloadable for local inference.
Supports tools such as retrieval, web search and a calculator, configured per run in the inference script.
Handles file references in questions, with documented formats and organization.
Includes the deep research benchmark evaluation used by the project.
The model is available on OpenRouter, so you can run inference with no GPUs at all.
Get the source code from GitHub and enter the project directory.
git clone https://github.com/Alibaba-NLP/DeepResearch.git
cd DeepResearchThe README's example uses conda with Python 3.10, then installs the requirements.
conda create -n react_infer_env python=3.10.0
conda activate react_infer_env
pip install -r requirements.txtEdit run_react_infer.sh (MODEL_PATH, DATASET, OUTPUT_PATH and any tool API keys), then run it.
# Edit run_react_infer.sh per README step 4, then run the script - see README step 5It is maintained by Alibaba's Tongyi team (Alibaba-NLP) on GitHub.
Yes - the project is Apache-2.0 licensed; model weights are provided under the terms on their download pages.
Local runs need no API key beyond any tool credentials you enable (for example a search API). The OpenRouter route requires an OpenRouter account and is pay-per-use.
Compared with hosted deep-research features in chatbots, Tongyi Deep Research gives you the open model and agent code to run, inspect and benchmark yourself.
Clone the repo, create the conda env, install requirements, download the weights, configure run_react_infer.sh and run your first question file.
Deep research that stays on your machine — local or cloud LLMs, 10+ search engines, encrypted and private.
Model-agnostic marketplace of agent skills with a harness-free canonical layer.
ESP32 iPhone Agent — let an AI model use a real iPhone with a $7 ESP32: AirPlay eyes, USB HID hands, no computer, no jailbreak
Snowflake's Agent World Model: synthesise 1,000 executable MCP tool-use environments for agentic RL.
Alibaba's in-page GUI agent: control any web interface with natural language, from one script tag
Jev Ultrafast — the fastest browser agent from browser-use: one decision-model request per action cycle drives a real browser, MIT
Другие проекты из категории «ИИ-агенты» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если Tongyi Deep Research вам не подходит.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Соседние проекты из категории «ИИ-агенты» в каталоге, ближайшие по числу звёзд к Tongyi Deep Research — продолжайте изучать полку по обе стороны от него.
Kortix - the open-source AI operating system: agents, skills and memory as files in a repo
NVIDIA's security scanner for AI agent skills - check before you install
Give it a GitHub issue and it tries to fix the code for you - the original SWE-bench agent
Devika - the open-source agentic software engineer that plans, browses and writes code
The veteran open-source framework for conversational AI that you host.
Open source agentic operating system
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.