LiveAgent
A fully functional AI Agent desktop client that supports Webui access and can be creatively customized and expanded!
What is LiveAgent?
A fully functional AI Agent desktop client that supports Webui access and can be creatively customized and expanded!
Overview of LiveAgent
What it is
LiveAgent is a local-first AI agent desktop client. It integrates large language model reasoning with local system tools, enabling the AI to operate the file system, run commands, and manage scheduled tasks. The Gateway component provides remote access and collaboration from any browser. The software is open source under the MIT license.
Key features
- Multi-model routing: supports Claude (Anthropic), Codex (OpenAI), and Gemini protocols with custom Base URL support for third-party compatible services.
- Local tool execution: includes full file-system capabilities (Read, Write, Edit, Delete, Glob, Grep), Bash and long-running process supervision, sub-agent delegation with worktree isolation, and tunnel exposure via TunnelManager.
- MCP & Skills ecosystem: bridges any stdio or HTTP MCP server for tool extension; supports Skills packages with progressive disclosure, on-demand loading, and the ClawHub ecosystem.
- Memory & automation: persistent memory using Markdown and SQLite FTS full-text search; scheduled tasks of types bash, http, and prompt, executed automatically in the background.
- Remote Gateway: a Go gateway using WebSocket and Protobuf provides a WebUI for remote control; disconnect recovery uses a bounded sequence window and desktop-side persistence.
- History compaction: dual-layer Segment and Summary Checkpoint persistence to maintain long conversation context.
- Internationalization: built-in i18n multi-language framework.
Use cases
LiveAgent is suitable for users who need an AI agent that can perform real system operations beyond chat, such as reading and writing files, running commands, and supervising long-running processes. It can be used for local development tasks, automated file management, and scheduled job execution. The remote Gateway enables controlling the desktop agent from a browser, useful for remote work or collaboration. The MCP and Skills ecosystem allows extending the agent with external tools and on-demand skill packages.
Getting started
Installers are built and published via GitHub Releases. System requirements: macOS (Intel x64 and Apple Silicon), Windows x64 (requires WebView2 runtime), Linux x86_64 (requires WebKitGTK 4.1). For macOS, download the DMG matching your chip and drag into Applications. For Windows, choose Setup wizard, MSI, or portable. For Linux, use AppImage, DEB, or RPM. To deploy the Gateway for remote access, use Docker:bashdocker pull ghcr.io/stack-cairn/liveagent-gateway:latestdocker run -d --name liveagent-gateway --restart unless-stopped -p 3000:8080 -v liveagent-gateway-data:/var/lib/liveagent -e LIVEAGENT_GATEWAY_TOKEN=your-token ghcr.io/stack-cairn/liveagent-gateway:latest
License and pricing
The license is MIT. The software is open source with no stated pricing.
Based on the project's public README and repository metadata.