Persistent chat automation, intelligent tool execution, scheduled task orchestration, and multi-channel message delivery. 26 LLM providers, 13+ channel backends, hybrid model routing, and permanent memory.
Linux (x64/ARM64), macOS (Intel/Apple Silicon), Windows (x64/ARM64)
# Auto-detects OS and architecture
curl -sSL https://guoqingbao.github.io/xbot/install.sh | bashnpm install -g @trusted-ai/xbotcargo install xbotgit clone https://github.com/guoqingbao/xbot.git
cd xbot
cargo install --path .From interactive CLI to always-on backend, xbot covers every AI agent use case.
Long-running backend with JSONL session persistence, configurable concurrency (default 3), and three operating modes: CLI, REPL, and backend service.
LLM-driven memory consolidation (auto-triggers at 75% context), dual-file MEMORY.md + HISTORY.md architecture, /memorize command, and topic-relevant retrieval.
File read/write/edit, directory listing, grep search, shell exec, web search/fetch, messaging, cron scheduling, and subagent spawning.
Main task on remote frontier API (OpenAI, DeepSeek), subagents fan out onto local models (Ollama, vLLM) for parallel execution — maximizing efficiency and cost control.
Slack, Telegram, Discord, Feishu, DingTalk, WeChat, WeCom, QQ, WhatsApp, Matrix, Email, MoChat, and more — unified message bus architecture.
Cloud APIs (OpenAI, Anthropic, DeepSeek, Gemini, Groq, Mistral) plus local inference (Ollama, vLLM) with automatic model routing.
MCP stdio protocol support for external tool servers (GitHub, databases, filesystems). Tools register as mcp_server_tool format.
Software engineering, research, GitHub/CI, scheduled ops, memory management, data analysis, project initialization, and more modular skills.
One-shot, interval, and cron expression scheduling with built-in heartbeat service (default 30-minute autonomous task review).
Cloud APIs, local inference, OAuth login — auto-detect routing, one config fits all models.
Pre-compiled binaries for every platform. Download and run.
| Platform | Architecture | Format | Download |
|---|---|---|---|
| Linux | x86_64 | .deb / .tar.gz | Download → |
| Linux | ARM64 | .deb / .tar.gz | Download → |
| macOS | Apple Silicon | .tar.gz | Download → |
| macOS | Intel x64 | .tar.gz | Download → |
| Windows | x86_64 | .zip | Download → |
| Windows | ARM64 | .zip | Download → |
Receive and reply across Slack, Telegram, Discord, WeChat simultaneously — unified management for all conversations.
File management, shell execution, GitHub CI/CD integration, and scheduled ops — fully autonomous execution.
Web search, data fetching, multi-source aggregation, and automated research report generation.
Guides, architecture docs, API reference, and operations manual.
Install, configure, and run for the first time
Operating modes, memory, skills, hybrid routing, MCP, gateway
Detailed credential and configuration guides for 13+ channels
Architecture, module structure, contributing, extending
CLI commands, env vars, gateway endpoints, admin API
Remote main model + local subagent routing