AideAgent: A Desktop AI Assistant with 28 Integrated Tools, Supporting Multi-Models and Local Knowledge Bases

What is AideAgent?

AideAgent is a powerful desktop AI assistant that supports multiple large models such as DeepSeek, Claude, GLM, Qwen, etc. It includes 28 built-in professional tools covering programming, search, file management, and other fields. It also supports local knowledge base RAG, MCP protocol extension, and a self-evolving memory system.

Source Information

• Original author/maintainer: quanzefeng
• Source platform: GitHub
• Project link: https://github.com/quanzefeng/AideAgent
• Release time: June 5, 2026

With the development of large language models, AI assistants have evolved into productivity tools, but most suffer from single-function limitations or difficulty integrating into existing workflows.

AideAgent addresses these issues: it not only supports multi-model switching but also has rich built-in tools, combined with local knowledge base and self-evolving memory, making it a truly all-round productivity assistant.

AideAgent supports 8 mainstream model providers, including DeepSeek, GLM (Zhipu), Qwen (Tongyi Qianwen), Claude (Anthropic), MiniMax, and local deployment options (llama.cpp, LM Studio, Ollama). Users can choose models based on task requirements, cost, and privacy.

It is also compatible with OpenAI and Anthropic API formats, allowing custom API addresses to easily access models not in the preset list.

AideAgent has 28 built-in tools divided into several categories:

• File & Code Operations: file_read/write/edit, grep, glob, lsp (code intelligence features)
• Command Execution: bash (cross-platform) with dangerous operation detection
• Web Search: Tavily API (recommended), built-in meta search engine (free), web_fetch (with SSRF protection)
• Git & GitHub Integration: git_diff/commit/branch, gh_pr/issue/repo
• Knowledge Base & Memory: kb_search/write, write_memory
• Task Management: TaskCreate/Update/List, TodoWrite
• Sub-Agent System: Agent (parallel execution of research tasks)

MCP Protocol Support

AideAgent supports the Model Context Protocol (MCP), which can connect external tools from local processes or remote services. It automatically detects local MCP configurations like Claude Code, enabling seamless integration of built-in and external tools.

Local Knowledge Base RAG

Integrated with Obsidian vault to enable local knowledge retrieval: hybrid search (FTS5 full-text + vector semantic + RRF fusion), supports MiniLM-L6 offline vector model and Ollama custom models, and relevant notes are automatically injected into the conversation context.

Self-Evolving Memory System

Memory is stored as Markdown files (with PREFERENCE/DECISION/KNOWLEDGE/REFERENCE tags). During conversations, relevant memory is automatically selected and injected with auto-deduplication. After the conversation ends, background analysis extracts and saves memory.

AideAgent is suitable for various scenarios:

• Personal Developers: Programming assistance, document query, project management
• Team Collaboration: Git workflow automation, code review assistance
• Researchers: Literature retrieval, experimental code writing, note management
• Enterprise Users: Internal knowledge base Q&A, code standard review, automated workflow building

AideAgent represents a new direction for desktop AI assistants. Its core advantages include openness (multi-model/protocol), privacy-first (local knowledge base/model), extensibility (MCP/Hook), intelligence (auto memory/search), and security (dangerous detection/protection).

It is recommended for developers who value data privacy, need flexible model selection, or want to integrate personal knowledge bases to try this tool. The project is open-source, and the community can contribute new features for continuous evolution.