llms.txt
# Heym
> Infrastructure for running and governing AI agents in production. Build agentic systems. Run them with confidence.
Heym is source-available, self-hosted infrastructure for running and governing AI agents in production. Teams use it to orchestrate agents, automate business processes, inspect every execution, and keep humans in control on their own infrastructure. It features a visual drag-and-drop canvas, multi-agent orchestration, built-in RAG pipelines, execution traces with per-run cost, evals, human-in-the-loop checkpoints, user-built workflow dashboards, MCP support, and a portable skills system.
Site search: https://heym.run/?s={query} (searches pages, templates, blog posts, and docs)
## Core Capabilities
- **Visual Workflow Editor**: Drag-and-drop canvas powered by Vue Flow with 65 node types across 7 categories (Triggers, AI, Logic, Data, Integrations, Automation, Utilities).
- **AI Assistant**: Describe what you want in natural language or voice, and the assistant generates nodes and edges and applies them to the canvas instantly.
- **User-Built Dashboards**: Build chart dashboards where every widget is powered by a hidden Heym workflow ending in a Chart Output node. Supports AI widget generation, fine-tuning, TTL caching, and drag-resize layouts.
- **Multi-Agent Orchestration**: One agent orchestrates named sub-agents and sub-workflows, all wired visually. Supports configurable reasoning effort, tool calling, and automatic model fallback.
- **Built-in RAG Pipeline**: Upload PDFs, Markdown, CSV, or JSON to managed QDrant vector stores. Wire a RAG node into any workflow for semantic search.
- **Decision Models**: Use Jev, Laya, or custom decision models in the [Decision node](https://heym.run/docs/nodes/decision-node). Ask typed questions, receive probabilities, and route with Condition or Switch. Supports AI-drafted questions and custom JSON request bodies.
- **MCP Support**: Connect to MCP servers (as a client) or expose workflows as an MCP server for Claude Desktop, Cursor, and other clients.
- **Skills System**: Portable capability bundles (SKILL.md + optional Python tools) that extend agent context and toolbox via drag-and-drop.
- **Human-in-the-Loop**: Pause execution for user approval before proceeding.
- **Guardrails**: Content filtering, NSFW protection, and safety checks at the node level.
- **Parallel DAG Execution**: Independent nodes run concurrently based on graph structure.
## Tech Stack
- **Frontend**: Vue.js 3, TypeScript, Vite + Bun, Vue Flow, Pinia, Tailwind CSS, Shadcn Vue
- **Backend**: Python 3.11+, FastAPI, UV package manager, Alembic migrations
- **Database & Infra**: PostgreSQL 16, SQLAlchemy 2.0, Redis, RabbitMQ, Docker, QDrant
- **AI & LLM**: OpenAI, Ollama (local), vLLM, Cohere, MCP protocol, HITL, Evals
- **Security**: JWT auth, bcrypt, Pydantic v2 validation, Fernet encryption
## Platform Operations
Enterprise identity, high availability, and the audit trail ship in the source-available product. Details: https://heym.run/enterprise
- **Single sign-on (SSO)**: OpenID Connect against your own identity provider. Keycloak, Okta, Entra ID, Auth0, and Google connect from a single issuer URL, with Authorization Code and PKCE, encrypted client secrets, just-in-time account provisioning, domain-restricted access, and password-free sign-in once SSO is verified. Reference: https://heym.run/docs/reference/sso
- **High availability and horizontal scaling**: Additional instances join as workers against the same PostgreSQL database and share workflow execution through weighted load balancing you control. Leader election keeps cron schedules, alerts, and crash recovery running through a failover, and every run records the instance that executed it. Reference: https://heym.run/docs/reference/cluster
- **Audit logging**: An always-on audit trail covering every privileged action with actor, target, and outcome, streamed alongside your container logs into any SIEM or log collector for retention and search. Reference: https://heym.run/docs/reference/audit
- **Access control**: Role-based team permissions, encrypted credential storage, and credentials shared by reference so team members use them without seeing the secret values.
- **Deployment and data residency**: Self-hosted on Docker Compose or Kubernetes inside your own VPC or on-premises perimeter, with local model inference through Ollama or vLLM.
## Node Types
All 65 node types across 7 categories (Triggers, AI, Logic, Data, Integrations, Automation, Utilities). Each has a documentation page under https://heym.run/docs/nodes.
Input, Cron, Telegram Trigger, WebSocket Trigger, File Upload Trigger, Output, LLM, Decision, AI Agent, Codex, OpenCode Go, Condition, Switch, Wait, Execute, HTTP, WebSocket Send, Sticky Note, Merge, Set, Converter, Code, JSON output mapper, HTML output mapper, Slack, Discord, Telegram, Slack Trigger, Discord Trigger, IMAP Trigger, Google Drive Trigger, Google Sheets Trigger, Send Email, Error Handler, Variable, Loop, Disable Node, Enable Node, Redis, RAG / Vector Store, Grist, GitHub, Linear, Jira, Google Sheets, Google Drive, DataTable, Throw Error, RabbitMQ, Crawler, Console Log, Playwright, Drive, Amazon S3, BigQuery, Supabase, ClickHouse, Notion, Sentry, MCP Call, Chart Output, Plugin, Plugin Trigger, Heym Trigger, Heym
## Documentation
Full documentation is on this site at https://heym.run/docs.
- **Getting Started** (https://heym.run/docs#getting-started): 6 pages. Install Heym, understand the canvas, and build a first workflow.
- **Node Reference** (https://heym.run/docs#nodes): 65 pages. Configuration, inputs, outputs, and expressions for every node type.
- **Dashboard Tabs** (https://heym.run/docs#tabs): 18 pages. What each dashboard tab manages and how permissions apply.
- **Reference** (https://heym.run/docs#reference): 44 pages. Expression DSL, agent architecture, guardrails, security, and platform internals.
Start here:
- [Core Concepts](https://heym.run/docs/getting-started/core-concepts)
- [Introduction](https://heym.run/docs/getting-started/introduction)
- [Migrate to Heym](https://heym.run/docs/getting-started/migrate-to-heym)
- [Quick Start](https://heym.run/docs/getting-started/quick-start)
- [Running & Deployment](https://heym.run/docs/getting-started/running-and-deployment)
- [Why Heym](https://heym.run/docs/getting-started/why-heym)
## Changelog
130 published releases, v0.0.131 being the latest (2026-10-17). The full history, including breaking changes and security fixes, is at https://heym.run/changelog. It is generated from the GitHub release tags at https://github.com/heymrun/heym/releases and refreshed hourly.
Most recent releases:
- **v0.0.131** (2026-10-17, https://heym.run/changelog#v0-0-131): Heym v0.0.131: 10 merged pull requests.
- **v0.0.130** (2026-10-16, https://heym.run/changelog#v0-0-130): Heym v0.0.130: 8 merged pull requests.
- **v0.0.129** (2026-10-12, https://heym.run/changelog#v0-0-129): Heym Work, the business interface for Heym, is now supported: people describe a job in Chat, Heym builds and test-runs the workflow and marks it Verified, and the team works from…
- **v0.0.128** (2026-10-07, https://heym.run/changelog#v0-0-128): Heym v0.0.128: 22 merged pull requests.
- **v0.0.127** (2026-10-06, https://heym.run/changelog#v0-0-127): Heym v0.0.127: 13 merged pull requests.
- **v0.0.126** (2026-10-05, https://heym.run/changelog#v0-0-126): Heym v0.0.126: 19 merged pull requests.
- **v0.0.125** (2026-10-04, https://heym.run/changelog#v0-0-125): Heym v0.0.125: 5 merged pull requests.
- **v0.0.124** (2026-10-03, https://heym.run/changelog#v0-0-124): Heym v0.0.124: 8 merged pull requests.
- **v0.0.123** (2026-10-02, https://heym.run/changelog#v0-0-123): Heym v0.0.123: 5 merged pull requests.
- **v0.0.122** (2026-10-01, https://heym.run/changelog#v0-0-122): Heym v0.0.122: 6 merged pull requests.
## Solutions
Heym helps AI, platform, and automation teams turn business processes into controlled, observable agentic systems, from company knowledge and human-reviewed operations to dashboards, observability, embedded backends, and evals. Overview: https://heym.run/solutions
- **Knowledge and Policy Systems** (https://heym.run/solutions/knowledge-systems): answers grounded in approved company knowledge with traceable sources and human review. Internal knowledge assistants, policy assistants, authenticated portals.
- **Human-in-the-Loop AI Operations** (https://heym.run/solutions/human-reviewed-operations): sensitive processes automated with human approval before important decisions or actions move forward. Legal, finance, compliance, content approval.
- **Multi-Agent Orchestration** (https://heym.run/solutions/multi-agent-research): specialist agents that collect information, analyze findings, and produce structured reports. Market research, document extraction, report generation.
- **Self-Healing Browser Automation** (https://heym.run/solutions/browser-automation): Playwright steps or sandboxed Playwright Python with AI steps that repair broken selectors, cookie and storage state auth with login fallback, screenshots, and network capture.
- **Team Skills and Skill Operations** (https://heym.run/solutions/team-skills): skills bundling instructions, Python tools, and assets that whole teams reuse, with history and revert, Heym Drive file support, and MCP exposure to outside tools.
- **Live Operational Dashboards** (https://heym.run/solutions/operational-dashboards): dashboards where every chart widget is a workflow ending in a Chart Output node, with AI widget generation, fine-tuning, and TTL caching.
- **AI Agent Observability** (https://heym.run/solutions/ai-observability): traces for every agent run with payloads, tool calls, timing breakdowns, token and cost rollups per model, threshold alerts, and OpenTelemetry export.
- **Embedded AI Backends** (https://heym.run/solutions/embedded-ai-backends): workflows exposed as JSON APIs with bearer token auth, MCP servers, or authenticated portals, with structured outputs and async dispatch.
- **Evaluation and Release Gating** (https://heym.run/solutions/ai-evaluation): blind multi-model comparison, evals, and guardrails that test prompts and agent outputs before changes ship.
## Platform Concepts
Heym's own definitions for the category it operates in. Each page states what the concept is, what a system has to do to qualify, and how Heym implements it.
- **AI agent runtime** (https://heym.run/ai-agent-runtime): An AI agent runtime is the system that executes agents in production. It decides when an agent runs, calls the tools the agent asks for, holds the run’s state when it pauses, records what every model and tool call received and returned, accounts for the cost, and enforces the limits on what the agent may do without a person. A framework describes an agent. A runtime is what actually runs it, over and over, on a schedule, in front of real data.
- **Self-hosted AI agents** (https://heym.run/self-hosted-ai-agents): Self-hosted AI agents run on infrastructure you control rather than on a vendor’s multi-tenant service. The agent logic, the credentials it uses, the documents it retrieves from, and the record of what it did all stay inside your own network boundary, and you decide which model providers it is allowed to reach. What you gain is control over the data path and the deployment. What you take on is the operation of the runtime itself.
- **AI agent governance** (https://heym.run/ai-agent-governance): AI agent governance is the set of controls that decide what an agent may do on its own, what it must stop and ask about, and what record it leaves behind either way. It covers four moments: what reaches the model, what the model is allowed to act on, who is allowed to change any of it, and what can be reconstructed afterwards. Governance that lives in a policy document is a description. Governance that lives on the execution path is a control.
- **MCP agent runtime** (https://heym.run/mcp-agent-runtime): The Model Context Protocol is a standard way for a model to discover and call tools. An MCP agent runtime is the system on the other side of that protocol: it connects agents to MCP servers, calls the tools those servers advertise, decides which of those calls are allowed to run unattended, records every one of them, and can publish its own workflows back as MCP servers for other clients to call.
## Comparisons
Independent, source-linked product comparisons. Each page states where the other product can fit well and links to official vendor documentation; verify current capabilities, plans, pricing, and licensing with the vendor.
- **Comparison library** (https://heym.run/compare)
- **Heym vs n8n** (https://heym.run/compare/n8n)
- **Heym vs Zapier** (https://heym.run/compare/zapier)
- **Heym vs Make** (https://heym.run/compare/make)
- **Heym vs Pipedream** (https://heym.run/compare/pipedream)
- **Heym vs Activepieces** (https://heym.run/compare/activepieces)
- **Heym vs Microsoft Power Automate** (https://heym.run/compare/power-automate)
- **Heym vs Dify** (https://heym.run/compare/dify)
- **Heym vs Flowise** (https://heym.run/compare/flowise)
- **Heym vs Langflow** (https://heym.run/compare/langflow)
- **Heym vs Node-RED** (https://heym.run/compare/node-red)
- **Heym vs Windmill** (https://heym.run/compare/windmill)
- **Heym vs Gumloop** (https://heym.run/compare/gumloop)
## Migration
Moving an existing workflow onto Heym runs through AI Convert (https://heym.run/convert): paste an export, watch it get rebuilt as native Heym nodes on a live canvas, then download and import. Exported files never contain credentials, so those are reattached in Heym afterwards. Guide index: https://heym.run/migrate
- **n8n to Heym** (https://heym.run/migrate/n8n-to-heym): Move n8n workflows to Heym. Paste an exported workflow JSON, watch AI Convert rebuild it on a live canvas, then download and import it into your own instance. Maps 12 source constructs to their Heym equivalents.
- **Make to Heym** (https://heym.run/migrate/make-to-heym): Move Make scenarios to Heym. Paste an exported blueprint, watch AI Convert rebuild every module on a live canvas, then run it on your own infrastructure. Maps 12 source constructs to their Heym equivalents.
- **Zapier to Heym** (https://heym.run/migrate/zapier-to-heym): Move Zaps to Heym. Describe the Zap or paste its steps, watch AI Convert rebuild it as a workflow on a live canvas, then run it on infrastructure you control. Maps 12 source constructs to their Heym equivalents.
- **Dify to Heym** (https://heym.run/migrate/dify-to-heym): Move Dify apps to Heym. Paste an exported workflow, watch AI Convert rebuild every node on a live canvas, then run it on infrastructure you control. Maps 12 source constructs to their Heym equivalents.
- **Flowise to Heym** (https://heym.run/migrate/flowise-to-heym): Move Flowise chatflows and agentflows to Heym. Paste an exported flow, watch AI Convert rebuild it on a live canvas, then run it on your own infrastructure. Maps 10 source constructs to their Heym equivalents.
- **Langflow to Heym** (https://heym.run/migrate/langflow-to-heym): Move Langflow flows to Heym. Paste an exported flow JSON, watch AI Convert rebuild each component on a live canvas, then run it on your own infrastructure. Maps 12 source constructs to their Heym equivalents.
- **Activepieces to Heym** (https://heym.run/migrate/activepieces-to-heym): Move Activepieces flows to Heym. Paste an exported flow, watch AI Convert rebuild each piece on a live canvas, then run it on infrastructure you control. Maps 11 source constructs to their Heym equivalents.
- **Pipedream to Heym** (https://heym.run/migrate/pipedream-to-heym): Move Pipedream workflows to Heym. Paste the workflow or describe it, watch AI Convert rebuild each step on a live canvas, then run it on your own servers. Maps 11 source constructs to their Heym equivalents.
- **Node-RED to Heym** (https://heym.run/migrate/node-red-to-heym): Move Node-RED flows to Heym. Paste an exported flow JSON, watch AI Convert rebuild each node on a live canvas, then run it on your own infrastructure. Maps 12 source constructs to their Heym equivalents.
- **Windmill to Heym** (https://heym.run/migrate/windmill-to-heym): Move Windmill flows to Heym. Paste an exported flow, watch AI Convert rebuild each step on a live canvas, then run it on infrastructure you already control. Maps 12 source constructs to their Heym equivalents.
- **sim.ai to Heym** (https://heym.run/migrate/sim-ai-to-heym): Move sim.ai workflows to Heym. Paste an exported workflow, watch AI Convert rebuild every block on a live canvas, then run it on your own infrastructure. Maps 13 source constructs to their Heym equivalents.
- **Huginn to Heym** (https://heym.run/migrate/huginn-to-heym): Move Huginn scenarios to Heym. Paste an exported scenario, watch AI Convert rebuild each agent as a node on a live canvas, then run it on your own server. Maps 13 source constructs to their Heym equivalents.
## Use Cases
- **Customer Support**: RAG-powered knowledge base, human-in-the-loop escalation, multi-language support, Slack/email integration.
- **Data Processing**: Web crawling with Playwright/FlareSolverr, PDF extraction, LLM-powered data transformation, S3 object storage, Redis/PostgreSQL/vector store output, and dashboard widgets.
- **Content Generation**: Multi-agent research and writing pipelines, quality checks with guardrails, scheduled generation via cron.
- **DevOps**: Webhook triggers for CI/CD, AI-powered log analysis, automated incident response with approval gates.
- **Sales & Marketing**: Lead enrichment with web crawling, LLM-generated personalized outreach, CRM integration via HTTP/webhooks, and dashboard widgets for conversion metrics.
## Getting Started
```bash
git clone https://github.com/heymrun/heym.git
cd heym && cp .env.example .env
./run.sh
Ready at http://localhost:4017
```
## Links
- [Website](https://heym.run)
- [Documentation](https://heym.run/docs)
- [Solutions](https://heym.run/solutions): knowledge systems, human-reviewed operations, and multi-agent research as controlled agentic workflows
- [Platform](https://heym.run/platform): platform overview, use cases, tech stack, and frequently asked questions
- [Enterprise](https://heym.run/enterprise): Heym Work for business teams, how it runs next to Heym Core on your infrastructure, the controls it builds on (single sign-on, roles, approvals, audit logs), and deployment support
- [Pricing](https://heym.run/pricing): free self-hosted Heym Core, Heym Work paid seats, Enterprise pricing, and professional services
- [Templates](https://heym.run/templates)
- [Blog](https://heym.run/blog)
- [Changelog](https://heym.run/changelog): every published release with breaking changes, security fixes, and the merged pull requests behind each version
- [Playground](https://heym.run/playground): replays of real Heym runs, step by step, no signup
- [AI Convert](https://heym.run/convert): convert n8n, Make, Zapier, Node-RED, Windmill, Pipedream, Huginn, Flowise, Dify, Langflow, sim.ai, or Activepieces workflows to Heym
- [AI agent runtime](https://heym.run/ai-agent-runtime): what an agent runtime has to do, and the four category pages defining Heym's terms
- [Migrate](https://heym.run/migrate): per-source migration guides mapping each construct to its Heym equivalent
- [MCP endpoint](https://heym.run/mcp): public MCP server with site summary and site-wide search tools
- [GitHub](https://github.com/heymrun/heym)
- [Discord](https://discord.gg/T2wXvuVdMX)
- [Enterprise email](mailto:enterprise@heym.run)
## License
Commons Clause + MIT. Free to use, modify, and distribute. Not for commercial resale.
