A side project we started in 2019 accidentally turned into an AIOS and an AI agent platform
A Brazilian team's 2019 side project evolved into a self-hostable AI agent platform supporting local LLMs.
A small, persistent team of developers in Brazil has been quietly building what they call an "AI Operating System" (AIOS) since 2019, long before the current generative AI boom. Their open-source project, General Bots, started as a vision to consolidate the fragmented tools companies use—chatbots, dashboards, documents, APIs, and AI models—into a single, cohesive platform where AI could operate internally. Over five years, they've assembled a comprehensive suite of features including AI agents that execute complex workflows, conversational interfaces over company data, knowledge bases with vector search for RAG (retrieval-augmented generation), automation pipelines, and omnichannel customer interaction tools.
The platform's architecture is fundamentally different from most contemporary AI SaaS startups, prioritizing privacy and control. It is entirely self-hostable and designed to avoid vendor lock-in, allowing organizations to run it with different AI model providers, including local LLMs like Llama 3. This makes it particularly compelling for governments, healthcare, finance, and other regulated industries where sending sensitive data to external cloud APIs is prohibited or risky. The team notes with pride that many features now touted as revolutionary in AI were already part of their experimental builds years ago, validating their early vision for an integrated, agentic enterprise AI environment.
- Open-source, self-hostable AIOS platform built since 2019, avoiding reliance on any single vendor like OpenAI or Anthropic.
- Combines AI agents, automation, document processing, vector search, and CRM tools into one integrated system for internal operations.
- Supports local LLMs and multiple providers, targeting regulated industries (gov, finance, health) that cannot use external AI services.
Why It Matters
It provides a viable, controlled path for enterprises to deploy AI agents and automation without compromising sensitive data to third-party clouds.