Heym is unclaimed —

Heym

Build agentic systems. Run them with confidence.·heym.run

Heym is a source-available, self-hosted AI-native workflow automation platform. It provides a visual drag-and-drop canvas for building intelligent automations with multi-agent orchestration, built-in retrieval-augmented generation (RAG), and Model Context Protocol (MCP) support. Workflows can include human-in-the-loop approval checkpoints, content guardrails, and parallel DAG execution. The platform runs entirely on the customer's own infrastructure via Docker Compose, a published container image, or Kubernetes.

What it's for

Customer support with RAG-powered knowledge base and human-in-the-loop escalationData processing via web crawling, PDF extraction, and LLM-powered data transformationContent generation with multi-agent research and writing pipelinesDevOps automation with webhook triggers for CI/CDInternal knowledge and policy assistants grounded in company knowledgeHuman-reviewed operations for legal, finance, compliance, and content approvalMulti-agent research and processing for market research and report generationSelf-healing browser automation with PlaywrightTeam skill bundling and reuse across an organizationLive operational dashboards built from workflow outputAI observability and cost governance with traces and token/cost rollupsEmbedded AI backends exposing workflows as JSON APIs or MCP serversEvaluation and release gating for prompts and agent outputs

Features 22

Agentic Kanban board where cards trigger workflow chains
AI Assistant for natural language and voice workflow generation
Auto Heal for Playwright browser automation (self-repairing selectors)
Automatic context compression for long-running agents
Built-in RAG pipeline backed by Qdrant or Postgres (pgvector)
Content guardrails with configurable sensitivity
DataTables with typed columns and CSV import/export
Drive for browsing and managing skill-generated files
Evaluation suites with LLM-as-Judge scoring
Human-in-the-loop approval checkpoints with public review links
LLM Batch API mode for high-volume, cost-efficient runs
LLM trace inspection with timing, tool calls, and cost breakdowns
MCP client and server support (bidirectional)
Multi-agent orchestration with named sub-agents and sub-workflows up to 5 levels deep
Parallel DAG execution
Persistent agent memory with graph-based knowledge extraction
Portable skills system (SKILL.md + Python tools) with AI Skill Builder
Portal: turn any workflow into a public chat UI
Teams with shared credentials, workflows, and vector stores
User-built chart dashboards powered by hidden workflows and a Chart Output node
Visual drag-and-drop workflow canvas (Vue Flow)
Visual schedule view for cron workflows

At a glance

free tierself-hostableAPI

TypeWeb app

DeploymentSelf-hosted

PlatformsWeb, Windows, macOS, Linux

ForAI teams, platform teams, automation teams, developers, DevOps engineers

CompanyHeym · 2026

Integrations

ElevenLabsCohereOpenAIAnthropicOllamaCerebrasGoogle GeminiOpenRouterSlackDiscordGmailOutlookGitHubJiraLinearCloudbackTelegramNotionQdrantClickHouseRedisSupabasePostgreSQLAmazon S3RabbitMQWebSocketDockerMCPGristGoogle SheetsGoogle DriveBigQueryPlaywrightSentryEjentum

Resources

Docs

Pricing

Heym is free to self-host with no licensing fees or usage limits. A cloud-hosted version is announced as 'Coming Soon' but pricing for it is not published.

Self-HostedFree

Free to use and self-host; licensed under Commons Clause + MIT, which permits use, modification, and distribution but not resale as a paid hosted service. A cloud-hosted version is marked 'Coming Soon' and not yet priced.

No licensing fees, no usage limits