The Heym AI agent platform

Build, execute, govern, observe, and evaluate AI agents on one self-hosted runtime. Heym is an AI workflow platform you run on your own infrastructure.

Product tour

Everything you need to run AI agents in production

See the problems Heym removes, how each capability works, and why it matters in practice.

How Heym Solves It
01 / 18

Overview

Problem to answer

The problem

Teams keep stitching together separate tools for work that should feel simple.

The Heym answer

Heym brings sixteen once-difficult jobs onto one visual canvas.

116 practical capabilities
2Source-available
3Self-hosted

Use the arrow keys to move through the tour

Why it matters

This tour starts with the problem, then shows the Heym answer. Each stop focuses on a real task teams already handle today, without burying the point under a long feature list.

Heym product demo: a drag-and-drop canvas where agentic AI workflow nodes such as LLM, Agent, RAG / Vector Store, HTTP, and Slack are connected into a running automation pipeline.

AI agent runtime

The runtime for production AI agents

A prototype becomes a production system when five things hold at once. Heym is an AI workflow platform that carries all five in one runtime you host yourself, so an agent that works in a demo keeps working when it runs a thousand times a day.

  1. Step 01

    Build

    Lay out agents, tools, and logic on a visual canvas, or describe the workflow in plain language and let the AI assistant draw it for you.

    • Visual workflow canvas
    • Real-time collaboration and voice
    • 68 node types
    • Natural language builder
    • AI-filled node fields
    • Expression DSL with live preview
    • Workflow Analyzer reports
    • Importable templates
    See the builder
  2. Step 02

    Execute

    Run those workflows at production scale, with branches that execute in parallel, retries on failure, and worker instances that share the load.

    • Parallel DAG execution
    • Schedule, webhook, and API triggers
    • Per-workflow concurrency limit and queue
    • Sub-workflows
    • Horizontal scaling across instances
    • Sandboxed code and browser runs
    Browse the nodes
  3. Step 03

    Govern

    Decide what an agent may do on its own and what waits for a person, then keep a record of every privileged action.

    • Human review steps
    • LLM guardrails
    • Role-based team permissions
    • Encrypted credential storage
    • Single sign-on and audit trail
    Platform and security
  4. Step 04

    Observe

    See what every model call and tool call did on every run, how long it took, and what it cost, then get told when one of those numbers crosses a line.

    • Execution traces with payloads
    • Token and USD cost per model
    • Workflow-powered dashboards
    • Threshold alerts
    • OpenTelemetry export
    Explore observability
  5. Step 05

    Evaluate

    Score prompts, models, and agent behavior against a repeatable suite so a change is measured before it reaches production.

    • Prompt suites across models
    • Agent suites: tool calls, latency, cost
    • Generated test data
    • Exact match, contains, and judge scoring
    • Separate judge credential
    • Suites built from Chat or MCP
    • Release gating
    Explore evaluation

The same runtime covers all of it. Self-hosted AI agents run on your own infrastructure with Docker or Kubernetes, your credentials and run data stay inside your environment, and the source is available under an MIT license with a Commons Clause.

Solutions

Solutions for teams putting agentic systems to work

Heym focuses on ten solution areas where controlled, observable AI operations matter most. Each one is backed by real workflows you can import and run.

Built with Modern Tech

Every component is chosen for performance, developer experience, and reliability.

Frontend

The frontend uses Vue.js 3 with TypeScript for a fast, type-safe development experience. Vite handles builds and Tailwind CSS takes care of styling.

Vue.js 3TypeScriptViteTailwind CSS

Backend

A Python backend powered by FastAPI delivers async performance for concurrent workflow executions. Alembic handles database schema migrations.

Python 3.11+FastAPIAlembic

Database & Infra

PostgreSQL stores workflows and execution history, while pgvector powers built-in vector search for RAG pipelines. Docker keeps local development consistent.

PostgreSQLSQLAlchemypgvectorDocker

Security & Auth

JWT tokens in HttpOnly cookies with refresh token rotation secure user sessions. Passwords use bcrypt hashing, credentials are encrypted at rest with Fernet, and MultiFernet rotates the encryption key without downtime.

JWTbcryptPydantic v2FernetMultiFernet

AI & LLM

Connect to OpenAI, Ollama for local models, vLLM for high-throughput inference, or expose workflows through the Model Context Protocol.

OpenAIOllamavLLMMCP

Developer Experience

First-party documentation lives inside Heym, alongside a native chat-with-docs experience. ESLint and Ruff keep the codebase healthy.

In-app docsChat with docsESLintRuff

Frequently Asked Questions

Answers to your common questions about Heym.

How is Heym different from workflow automation tools?

Heym is built to run and govern AI agents in production, and it automates the business processes around them in the same runtime. Multi-agent orchestration, human-in-the-loop checkpoints, node-level guardrails, built-in RAG, MCP, evals, and execution traces with per-run cost all live in one self-hosted system, so you can build agentic systems and run them with confidence.

Is Heym free to use?

Yes. Heym is source-available and free to self-host under the Commons Clause + MIT licensing model. You can deploy it on your own infrastructure without licensing fees. Commercial licensing and enterprise support are available for organizations that need those additional rights and services.

Can I self-host Heym?

Absolutely. Heym is designed for self-hosting. You can boot it locally with run.sh, deploy it with Docker Compose, use the published container image, or run it on Kubernetes. Your data stays in your infrastructure, giving you complete control and data sovereignty.

What LLM providers does Heym support?

Heym supports OpenAI (GPT-4, GPT-3.5), Ollama (for local LLMs), vLLM, Cohere, and any OpenAI-compatible API. You can also configure custom endpoints for your own models.

How does the AI Assistant work?

The AI Assistant lets you describe what you want in natural language or use voice input. It analyzes your request, generates appropriate nodes and edges, and applies them directly to the canvas. When your workflow already includes Agent skills, Heym keeps the builder context focused by sending only each skill's SKILL.md instead of large Python or binary attachments.

What is MCP support?

MCP (Model Context Protocol) is a standardized protocol for connecting AI assistants to tools and data. Heym supports MCP in two ways: as a client (Agent nodes can connect to any MCP server and gain all its tools), and as a server (your Heym workflows can be exposed as an MCP server for Claude Desktop, Cursor, and other clients). The heym_chat tool also brings Heym’s Chat to those clients, so one message can build and test a workflow or create, edit, and run eval suites. Heym also serves the same agent over A2A, the Agent2Agent protocol, when you turn it on.

How does RAG work in Heym?

Heym has built-in vector store management backed by either Qdrant or Postgres (pgvector) inside Heym itself, so you can run RAG with no external vector service. You can upload PDFs, Markdown, CSV, JSON, and other document types. The RAG node performs semantic search across your documents and returns relevant context that flows directly into your LLM or Agent nodes. Turn on hybrid search to add BM25 keyword matching, so exact terms such as product codes or error IDs are found too.

What are skills?

Skills are portable capability bundles. Each skill consists of a SKILL.md instruction file plus optional Python tools. You can drag and drop a .zip or .md file onto an Agent node, or use the built-in AI Skill Builder to draft and revise skills with live file previews. Skills enable code reuse and sharing across workflows and teams.

How can I contribute to Heym?

We love contributions! You can contribute by reporting bugs, suggesting features, submitting pull requests, improving documentation, or sharing your workflows with the community. Join our Discord to connect with other contributors and get started.

Is there commercial support available?

Yes! Enterprise licensing includes professional support with SLA guarantees. Contact us at enterprise@heym.run for information about commercial use, custom development, and priority support.

See enterprise licensing and support
What is the Commons Clause license?

The Commons Clause is a license condition that restricts selling the software or offering it as a paid service. You are free to use, modify, and distribute the software, but you may not resell it. This keeps Heym transparent and accessible while preventing commercial redistribution.

How do I get started?

Getting started is easy. Clone the repository, copy the example environment file, and run the setup script. The script starts PostgreSQL, runs database migrations, and starts both the frontend and backend servers. If you prefer a prebuilt install, you can also use the published container image. Open your browser at the configured port (default: 4017) and you're ready to build.

Still have questions?

Ask on Discord