Auto Deep Research

Report Abuse

Basic Information

Auto-Deep-Research is an open-source, ready-to-run personal AI assistant that provides a cost-effective alternative to proprietary Deep Research offerings. Built on the AutoAgent framework, the repository packs a fully automated research-assistant experience intended for users who want to self-host an agent with minimal setup. It supplies installation instructions for Conda and Docker, a one-click launch command, and expects users to provide API keys for the LLM providers they choose. The project supports both function-calling and non-function-calling interactions, handles file uploads to enrich data interaction, and reports competitive performance on the GAIA benchmark. It is positioned as a practical, extensible agent app demonstrating AutoAgent capabilities while remaining configurable through environment variables and command options.

Links

Categorization

App Details

Features
High-performance results reported on the GAIA benchmark. Universal LLM compatibility with many providers listed in the README, including OpenAI, Anthropic, Mistral, Gemini, Huggingface, Groq, DeepSeek and OpenRouter. Support for both function-calling and non-function-calling LLMs. Cost-efficient operation by using user-supplied pay-as-you-go API keys. File upload handling for richer data-driven interactions. One-click launch command with zero-configuration defaults. Docker containerization with configurable container name and port. Command options for specifying completion model, debug mode, API base URL and function-calling toggle. Optional browser cookie import to improve website access. Documentation and community channels noted and a web GUI under development.
Use Cases
Auto-Deep-Research enables individuals and teams to run a personal assistant locally or in a containerized environment, avoiding high-cost subscriptions by using their own LLM API keys. The one-click launch and packaged installation steps reduce setup effort while Docker ensures a consistent runtime. Broad provider support gives users flexibility to select models that balance cost and performance. File upload and cookie import features expand the assistant"s ability to interact with user data and websites for research tasks. Configurable options let users tune behavior and debugging. The project also serves as a practical example of building agent apps on the AutoAgent framework and provides community and documentation touchpoints for support and extension.

Please fill the required fields*