Heym

Heym is infrastructure for building, running and governing AI agents in production, with a visual editor for assembling workflows. You can describe a workflow in natural language and let the assistant build it, or wire it yourself from a library of built in nodes. The aim is to orchestrate agents, automate business processes, inspect every execution and keep humans in control. Capabilities listed include LLM nodes for text and vision, agent nodes with tool calling and Python tools, retrieval over Qdrant or built in Postgres with pgvector, MCP client and server support, human approval checkpoints, content guardrails, parallel execution, a skills system for portable agent capabilities, and Playwright browser automation with auto heal. Integrations cover OpenAI, Anthropic, Google Gemini, Ollama, OpenRouter, Slack, Discord, Gmail, Outlook, GitHub, Jira, Telegram, Redis, RabbitMQ and any HTTP API. Tracing and evals are built in. It is aimed at AI, platform and automation teams, including those in regulated sectors such as financial services, healthcare and government. Heym is source available and fully self hosted, deployable with Docker Compose or Kubernetes on your own infrastructure so data and runs stay under your control. A hosted cloud version is marked as coming soon, and the site offers demo requests and an interactive demo. No price list is published; structured data on the site shows a zero dollar entry for the self hosted option.