Heym
A self-hosted runtime for building, orchestrating, running, observing, evaluating, and controlling agentic systems on a company's own infrastructure, composed visually or generated with AI.
Heym is a self-hosted runtime and platform for building, orchestrating, running, observing, evaluating, and controlling agentic systems. A workflow defines how agents, deterministic steps, tools, and data work together; the runtime executes that graph, coordinates dependent and parallel work, records execution history, and supports human review checkpoints, connecting to models, tools, MCP servers, company knowledge, and internal systems.
Its visual canvas is one interface into Heym, used to build, inspect, debug, and understand a system, but the same workflows also run through APIs, schedules, event triggers, MCP clients, and a chat interface called Portal. Stages of operating a system in Heym include composing it visually or generating it with AI; coordinating agents, sub-agents, and tools with explicit data flow and branching; running it from an API, schedule, or trigger with configurable retries; inspecting execution history, traces, and model costs, including exporting OpenTelemetry traces; testing prompts and model responses against expected outputs; pausing execution at review checkpoints for a human decision; and exposing workflows through REST, SSE streaming, MCP tools, or chat.
It is installed locally with a single clone-and-run script.