I built LX Engine to scratch my own itch when orchestrating multiple AI agents without constantly rewriting boilerplate code. It's an open-source execution engine that lets you define your entire AI workflow in a single JSON object—no more tedious manual pipeline setups. Just pass your agent pipeline as an input, and the engine handles orchestration automatically, returning either detailed step-by-step results or just the final output. It also comes with 20+ built-in agents (web retrieval, PDF creation, local embeddings, sandboxed code execution, etc.) accessible individually through dedicated endpoints. Still early days, feedback very welcome!
Show HN: Open-Source API-Based Engine to Easily Orchestrate AI Agents | Heykuki News