AQBot-Desktop/AQBot

What it solves

AQBot provides a local-first desktop AI workbench that centralizes multiple AI service providers, agents, knowledge bases, and tools into a single interface. It solves the problem of fragmented AI workflows by allowing users to manage various LLM APIs, local files, and external agent processes while keeping application data and user files under local control.

How it works

AQBot acts as a unified hub that connects to various AI providers (like OpenAI, Claude, Gemini, and DeepSeek) and supports the Model Context Protocol (MCP) for extending capabilities. It features two distinct agent modes: a built-in mode using service provider APIs and a dedicated ACP (Agent Client Protocol) workbench for running external programming agents. It also integrates a local knowledge base using sqlite-vec for private document indexing and provides a local API gateway to expose AI interfaces to other local clients.

Who it’s for

It is designed for power users and developers who want a centralized, privacy-focused desktop environment to manage multiple AI models, automate tasks via agents, and organize their AI-driven knowledge and coding workflows.

Highlights

• Unified AI Hub: Connects to multiple providers and OpenAI-compatible endpoints with custom proxy and header support.
• Dual Agent System: Supports both API-based agents and external ACP-compatible programming agents (e.g., Codex, Cline).
• Local-First Privacy: Stores application state and API keys (AES-256 encrypted) locally, ensuring user data stays on the machine.
• Extensible Tooling: Integrates MCP servers for tool use and includes a built-in local API gateway for other CLI tools.
• Knowledge Management: Features a local RAG system with sqlite-vec and integrated web search (Tavily, Exa, etc.).
• Rich Content Rendering: Supports LaTeX, Mermaid, D2 diagrams, and an Artifacts panel for code and reports.