OpenAgent

// Overview

OpenAgent is a self-hostable personal AI assistant that puts three things behind one web UI: any model provider you care to connect, a knowledge base built from your own documents, and autonomous agent loops that can browse the web, run code and call MCP-compatible tools on your behalf. It is written in Go and distributed as a single binary, so getting it running is a download and a port rather than a stack to assemble.

The agent side covers browser-use (navigate, click, fill forms, scrape, screenshot), web search and fetch, shell execution, Office file automation for Word, Excel and PowerPoint, and MCP integration over SSE, stdio or streamable HTTP. Every tool call is shown with its arguments and return value, step by step, which makes the loop auditable rather than a black box.

Retrieval is first-class: documents are chunked, embedded and indexed on upload, semantic search runs before each response, embeddings are pluggable across OpenAI, Azure, Gemini, Qwen, Cohere, Jina, HuggingFace and local models, and knowledge is organised into isolated stores you assign per chat or per application. A BPMN-style visual workflow builder adds conditional and parallel execution plus scheduled jobs, with usage analytics tracking token spend per provider, model and user. The project is Apache-2.0 licensed.

// What it does

• Single Go binary for Linux, macOS and Windows on x86_64 and arm64 — Windows runs natively with no WSL or Docker
• 30+ model providers, switchable per conversation without code changes
• Agent loops with browser-use, web search and fetch, shell execution and Office file automation
• MCP integration over SSE, stdio or streamable HTTP, with every tool call and its arguments shown step by step
• Built-in RAG: automatic ingestion and indexing, pluggable embedding providers, and isolated knowledge stores per chat or app
• BPMN-style visual workflow builder with conditional branching, parallel execution and scheduled runs
• Usage analytics for token consumption and cost per provider, model and user

// Getting started

The installer downloads the latest release and starts OpenAgent on port 14000. Prebuilt binaries cover Linux, macOS and Windows on x86_64 and arm64.

curl -fsSL https://raw.githubusercontent.com/the-open-agent/openagent/master/scripts/install.sh | bash

irm https://raw.githubusercontent.com/the-open-agent/openagent/master/scripts/install.ps1 | iex

http://localhost:14000

docker-compose up

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Run a private ChatGPT-style assistant on your own hardware with your documents as its knowledge base
• Let an agent complete a task end to end — search, drive a browser, run a script, write a spreadsheet — with every tool call visible
• Give a small team one place to connect many providers and track token cost per model and per user
• Schedule recurring agent jobs through the workflow builder instead of writing cron glue

// How OpenAgent compares

OpenAgent alongside other open-source assistants & chatbots tools AI/TLDR tracks, ranked by GitHub stars.