Hey Product Hunt! 👋 I'm Tim, founder of Synthetic AI. When we set out to put real AI into our IoT platform, the chat was the easy part. The hard part was everything underneath — the infrastructure that makes AI reliable on real-world data: retrieval, knowledge graphs, tools, memory, permissions, and a record of every action it takes. That took us six months to build properly. And it turned out to be far bigger than a feature — big enough to be a product of its own. So we open-sourced all of it. Synthetic Brew is a self-hosted AI agent runtime: - Describe the agent you need in plain English — the AI Builder configures the agents, tools, memory, and workflows (and yes, the Builder itself runs on Synthetic Brew) - Grounded with vector RAG + typed knowledge graphs, so it doesn't make things up - MCP tools, multi-agent ReAct orchestration, REST API + SSE streaming, embeddable chat widget - BYOK — OpenAI, Anthropic, Google, Groq, DeepSeek, or Ollama - One Docker command, your infrastructure, your data. No lock-in, no per-token markup. Honest license note: the Community Edition is BSL 1.1 (converts to Apache 2.0 after 4 years) — free forever to self-host, unlimited agents. I'll be here all day — ask me anything about the architecture, the knowledge-graph approach, or the six months of mistakes we made so you don't have to.

Synthetic Brew

Open-source AI agent runtime — describe it in plain English

Open-source AI agent runtime — describe it in plain English