Natural language workflow generation
Describe a workflow in plain text or voice and the assistant generates nodes and edges on the canvas, while still allowing manual editing afterward.
Heym is a source-available, self-hosted AI workflow automation platform for building and running agentic workflows with a visual editor. It helps automation, platform, and AI teams orchestrate agents, RAG, approvals, and execution tracing on infrastructure they control.
Heym is an agentic AI workflow automation platform that combines a visual editor with runtime execution, traces, and review checkpoints. Users can build workflows from natural language or wire them manually from built-in nodes and integrations, then run them on their own infrastructure.
It is aimed at automation, platform, and AI teams that want to coordinate agents, apply RAG, connect to external tools, and keep humans in the loop for critical steps. The platform is source-available and self-hosted, with support for Docker Compose and Kubernetes deployment.
Describe a workflow in plain text or voice and the assistant generates nodes and edges on the canvas, while still allowing manual editing afterward.
Build automations in a drag-and-drop canvas with built-in node types, version history, keyboard shortcuts, data pinning, and sub-workflow extraction.
Use one orchestrator agent to delegate work to named sub-agents and sub-workflows, with support for tool calling, Python tools, and nesting.
Upload PDFs, Markdown, CSV, or JSON to managed vector stores and run semantic search against Qdrant or built-in Postgres (pgvector), with optional reranking.
Add approval checkpoints, inspect every run with traces and evaluations, and monitor agent calls, tool calls, latency, and outcomes.
Run Heym on your own infrastructure using Docker Compose, a container image, or Kubernetes, with source-available code for review and customization.
Classify incoming requests, route them to specialist agents, and add a human approval step before any action that should be reviewed.
Combine RAG with LLM or agent nodes to search internal documents and feed retrieved context into downstream workflow steps.
Coordinate sub-agents, tools, and sub-workflows for tasks that need branching logic, sequential execution, and shared context.
Build dashboards where each chart is driven by its own workflow, making it possible to fetch, transform, and summarize operational data.
Use Playwright automation for web interactions that need page actions, with traces and review points available for inspection.
No. The site says workflows can be described in natural language and generated automatically, but users can also edit the canvas manually and work with built-in nodes.
Yes. Heym is described as source-available and fully self-hosted, with deployment options including Docker Compose and Kubernetes.
The platform includes multi-agent orchestration, RAG, MCP client and server support, human approval checkpoints, traces, evaluations, dashboards, and agent memory.
Yes. The site mentions integrations with tools and services such as Slack, Gmail, GitHub, Jira, Notion, PostgreSQL, Redis, RabbitMQ, Google Sheets, Google Drive, BigQuery, S3, and HTTP APIs, among others.
Not in the provided sources. The pricing URL returned a page not found response, so pricing and tiers could not be verified.
Traffic data is for reference only.
mindflow.io
Mindflow is a no-code agentic automation platform for enterprise IT and cybersecurity teams.
twin.so
Twin is a no-code AI agent platform for automating app connections, web browsing, scheduled workflows, and recurring team tasks.
zaro.ai
Zaro is an AI workspace for building internal apps, agents, and workflows on company data, with shared organisational credits, a free trial, and paid plans for higher usage.
taskagi.net
No-code AI platform for building and deploying agentic workflows
activepieces.com
AI-first automation platform for workflows and AI agents across business apps.
chatinsight.ai
Enterprise AI platform for automation, collaboration, and governed human-AI workflows