Heym
Heym is a self-hosted AI agent platform for building, running, and tracing agentic workflows on your own infrastructure.
Heym is the pick when agent workloads touch regulated data or your security team wants to read the source before it runs in production. Traces, evals, approvals, SSO, and audit logs sit in Heym Core rather than behind an enterprise tier, and pricing is per paid seat on Heym Work — not per run or per token, so a heavy execution load doesn't inflate your bill. The catch is that uptime, upgrades, and database scaling are yours to run.
Verified 3d ago · liveness 81/100 · cite: rightaichoice.com/tools/heym
- Engineering teams building production agentic workflows that must stay on their own infrastructure
- Regulated verticals — financial services, healthcare, government — needing audit trails and data residency
- Platform teams that want traces, evals, guardrails, RAG, MCP, and multi-agent orchestration in one self-hosted stack
- Business teams that need to run and approve agents without learning a canvas, on top of their platform team's Heym Core
- Non-technical users who want a managed, no-setup automation tool with vendor-hosted infrastructure
- Teams with nobody to own upgrades, database scaling, and infrastructure monitoring
- Anyone planning to resell hosted workflow automation — the Commons Clause forbids it
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Heym if you have no one to own PostgreSQL, container upgrades, and monitoring, or if you need a vendor SLA and 24/7 managed support rather than infrastructure you operate yourself.
Your model API bill is the real cost: Heym charges no licensing fee, but every Agent, LLM, and RAG node call is billed by whichever provider credential you configure.
Heym publishes a $0 Self-Hosted tier with no licensing fees and no usage limits, plus a Custom Enterprise tier for regulated rollouts — so your spend is infrastructure and model API calls rather than seats. That undercuts per-task or per-seat managed automation subscriptions, but it sits above tools where the vendor eats the hosting and ops burden.
In short
Heym — Heym is a self-hosted AI agent platform for building, running, and tracing agentic workflows on your own infrastructure. Best for Engineering teams building production agentic workflows that must stay on their own infrastructure, Regulated verticals — financial services, healthcare, government — needing audit trails and data residency, Platform teams that want traces, evals, guardrails, RAG, MCP, and multi-agent orchestration in one self-hosted stack. Free to start; paid plans from €99/mo.
What's new in Heym
Checked yesterdayAcross the latest 5 updates: 1 feature update, 3 changelog entries and 1 news mention.
v0.0.128
22 merged pull requests including an impl/run-to-node feature, Qdrant credential merge fixes, case-variant duplicate rejection at registration, and broad frontend test coverage.
v0.0.127 — Heym Work integration channel and admin tab
Adds a Heym Work integration channel with an admin tab, allows team-shared Drive attachments in the Send Email node, and generalizes credential merging across RabbitMQ, S3, and IMAP connectors.
Self-Hosted AI Agent Templates: 6 You Can Watch Run
Heym publishes six self-hosted agent templates with runnable setups, aimed at teams that want observable agents instead of demo code.
v0.0.118 — shared resources fix for multi-team users
Three merged PRs: shared resource resolution for users in multiple teams, deferred cluster allowDownstream finalization, and allow-downstream join masking.
v0.0.117 — cluster execution ownership handed to workers
Four merged PRs including handoff of cluster execution ownership to workers and a fix to use the highest share per node.
What people actually say about Heym — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
62 mentions across 5 sources (YouTube, Product Hunt, App Store, GitHub, Lemmy) · researched Aug 14, 2026.
Average across the 5 sources that answered — each source counts once, not each post.
- +All-in-one: agents, RAG, MCP, HITL, traces, evals in a single platform
- +Self-hosted deployment gives full control over data and infrastructure
- +Human-in-the-loop checkpoints make workflows production-ready
- +Visual builder with natural language workflow creation is intuitive
- +Built-in RAG with Qdrant or pgvector, plus document upload
- −Setup and configuration require DevOps skills; not beginner-friendly
- −License (Commons Clause + MIT) limits commercial reuse
- −Long-running agent reliability unproven; questions about context loss
- −Documentation may not cover all edge cases for complex workflows
- −Small user base means fewer community tutorials and shared solutions
- • Self-hosting requires your own infrastructure costs (servers, storage, LLM API usage)
- • No transparent pricing for paid tiers; likely usage-based costs add up
- • Potential costs for maintaining Kubernetes cluster if scaling
Viability Score
How well maintained and how widely used is Heym? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Visual canvas that executes branches, tool calls, agent memory, and HITL checkpoints as one live graph
- AI assistant that generates nodes, edges, and MCP tools from a natural-language workflow description
- Voice input and streaming responses inside the AI assistant
- Multi-agent orchestration with sub-agents, team skills, and sub-workflows from one orchestrator
- MCP client and server integration for connecting models to external tools
- RAG semantic search over Qdrant or built-in Postgres with pgvector
- LLM nodes for text generation and vision tasks
- Agent nodes with tool calling and custom Python tools
- Human-in-the-loop approval checkpoints before critical actions
- Content guardrails across input, output, action, and human review
- Parallel DAG execution of independent branches
- Playwright browser automation with auto-heal for broken selectors
- Execution traces, evals, and cost evaluation with release gating
- Live operational dashboards, monitoring, and alerting
- Heym Work team layer: describe a job, get a tested workflow, approve runs in pages built for business teams
About Heym
Heym is infrastructure for running and governing AI agents in production, paired with a visual editor for building them. It comes in two layers: Heym Core, the source-available runtime where engineers design and debug agentic systems, and Heym Work, a commercial team layer where business teams describe a job in plain language, get a tested workflow, and approve what matters before it goes out. The split matters — the platform team keeps sign-in, roles, and the audit trail, while the people using the agents never touch the canvas. Building happens either way you like. Describe a workflow in natural language and the assistant generates nodes, edges, and MCP tools on the canvas, with voice input and streaming responses built in; or wire it by hand from the node library. The canvas is a runtime, not a form builder: branches, tool calls, agent memory, and human-in-the-loop checkpoints execute as one visible graph, traced step by step. The capability set is aimed at production agent systems rather than demos. LLM nodes handle text generation and vision, Agent nodes do tool calling with Python tools, RAG runs semantic search over Qdrant or built-in Postgres with pgvector, and MCP client/server integration connects models to external tools. Multi-agent orchestration coordinates sub-agents, team skills, and sub-workflows from one orchestrator, with independent branches executing in parallel. Playwright browser automation ships with auto-heal for broken selectors, and content guardrails cover input, output, action, and human review. Deployment is deliberately unglamorous: self-host with Docker Compose or Kubernetes on your own infrastructure, bring your own AI providers, and keep your data and runs where you control them. Heym charges no execution fees, no compute fees, and no AI token markup — infrastructure and third-party API costs stay yours. That pricing model is the main thing separating it from managed agent platforms, and the tradeoff is that you own the uptime.
Behind the Verdict
The decision with Heym usually comes down to one question: who is allowed to see the data? If your agents touch claims records, patient data, or anything a regulator asks about, the self-hosted model removes a whole category of procurement conversation. You clone the repo, deploy with Docker Compose or Kubernetes, and your runs never leave infrastructure you own. Where it earns its place is the runtime-plus-governance combination. Traces, evals, cost evaluation, release gating, monitoring, and human-in-the-loop checkpoints are part of the product, not bolted on after launch. That is a different posture from the hosted agent builders, where observability tends to be a separate purchase. We'd reach for Heym Work specifically when the people running the agents aren't engineers. A support lead describes the job, Heym builds and tests the workflow before it's saved, and a run pauses before an email goes out or a record changes. That approval step is the part most automation tools leave to you. Pricing is refreshingly plain. Heym Core is free and source-available under Commons Clause plus MIT — free to use, modify, and distribute, but not to resell as a paid service. Heym Work runs €99 per paid seat per month (or €990 per seat billed annually), minimum five paid seats, with new monthly customers getting 50% off the first three months. End users of your workflows don't count as seats; a paid seat is someone who builds, configures, administers, or observes systems. No execution fees, no compute fees from Heym, no AI token markup. Where it bites: the release cadence. The project shipped 127 releases between April 27 and October 7, 2026, and the recent line is all about distributed execution — v0.0.117 handed cluster execution ownership to workers, v0.0.118 fixed
Researching Heym? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Heym actually fits — and what changes day-one when you adopt it.
Clones the Heym repo, runs the Docker Compose setup script, connects an Anthropic credential, and uses the AI assistant to generate a support-triage workflow that routes inbound email to specialist agents.
Outcome: A running graph in the browser at localhost within an afternoon, with traces and token cost visible per run and a HITL checkpoint before any customer reply is sent.
Builds a knowledge-and-policy system where a compliance checker node reads a centralized knowledge base and a guardrail layer blocks any agent action outside approved boundaries.
Outcome: Every agent execution is traced and inspectable, and policy coverage is reported on a dashboard the team can hand to auditors.
Runs existing n8n exports through the AI Convert tool, previews the rebuilt workflows on a live canvas, then imports them alongside new Playwright browser automations with auto-heal.
Outcome: Legacy automations run on self-hosted infrastructure with parallel DAG execution, and broken selectors repair themselves instead of paging someone at night.
Use Cases
- Build a multi-agent customer support system that answers tickets, routes complex issues to specialist agents, and drafts responses with human approval before sending.
- Create a RAG pipeline that ingests PDF and Markdown documents into Qdrant or Postgres with pgvector and serves semantic search with reranking.
- Automate social media or inbox monitoring by triggering workflows from Slack, Telegram, or an IMAP inbox and posting responses via webhooks.
- Route LLM calls through a decision model to cut cost and latency instead of hardcoding one provider per step.
- Deploy a self-hosted AI assistant that retrieves company knowledge and performs actions such as updating a CRM via MCP tools.
- Run knowledge-and-policy systems that feed compliance checkers and policy engines across an AI stack, keeping every agent inside approved boundaries.
- Build live operational dashboards where each widget runs a hidden Heym workflow to aggregate and render data.
- Migrate existing n8n, Make, Zapier, or Dify workflows into Heym with the AI Convert tool, preview them on a live canvas, then import.
Models Under the Hood
as of 2026-10-17
Limitations
- Heym is self-hosted: you run it on your own infrastructure, and infrastructure plus third-party AI/API costs remain with you — Heym itself charges no execution, compute, or token-markup fees.
- Building workflows requires familiarity with prompts, tool calling, RAG, and MCP, since agents are wired from a large library of nodes on a visual canvas (the AI assistant can generate them from natural language).
- Releases ship very rapidly (latest v0.0.130), so tracking upgrades is ongoing.
- Execution depends entirely on whatever LLM credentials you configure.
as of 2026-10-09
Verification history
We have re-verified Heym 10 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 10 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Heym tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Heym Core
$0
Heym Work
€99/paid seat/month
Enterprise
Custom
Ideal for
Financial services, healthcare, and government teams running business-critical agentic systems that need policy engines, audit trails, and data residency on their own infrastructure.
What this tier adds
Adds enterprise rollout support on top of self-hosting, with knowledge and policy systems for regulated industries and human approval before critical actions.
Where the pricing makes sense
The company stage and team size where Heym's pricing actually pencils out — and where peers do it cheaper.
Heym publishes a $0 Self-Hosted tier with no licensing fees and no usage limits, plus a Custom Enterprise tier for regulated rollouts — so your spend is infrastructure and model API calls rather than seats. That undercuts per-task or per-seat managed automation subscriptions, but it sits above tools where the vendor eats the hosting and ops burden.
Setup time & first value
How long it actually takes to get something useful out of Heym — broken out by persona, not the marketing-page minute.
Developers: the install video walks a self-hosted setup end to end and promises a local instance in under two minutes once model credentials exist. Teams: first production workflow typically lands in a day or two, then Postgres or Qdrant provisioning and Kubernetes rollout add real time. Regulated deployments are measured in weeks because knowledge and policy configuration is the bulk of the work.
Switching to or from Heym
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From n8n: export the workflow and run it through the AI Convert tool, preview it on a live canvas, then import.
- →From Make: convert the scenario with AI Convert, then rebuild any custom modules as Heym nodes.
- →From Zapier: import the zap via AI Convert for the linear portions and rewire multi-step branches by hand.
- →From Dify or Langflow: use AI Convert, then reconnect prompts and retrievers to Heym Agent and RAG nodes.
- →From hand-rolled Python scripts: rebuild the logic as a workflow with triggers, Agent nodes, and custom Python tools.
- ↗To a managed automation platform: export logic node by node, since Heym workflows and their traces are not portable to closed platforms.
- ↗To Langflow or Flowise: rebuild prompt and retrieval chains manually, as there is no documented reverse converter.
- ↗INSUFFICIENT_DATA
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Heym”, and we withheld 6: 6 could not be judged, because “Heym” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Heym.
Official links
Tools that pair well with Heym
Common stack mates teams adopt alongside Heym, with the specific reason each pairing earns its keep.
Stack AI
StackAI is an enterprise AI agent factory for building, deploying, and governing agentic workflows.
Notable
Notable is a healthcare AI agent platform that automates patient access, revenue cycle, care operations, and contact center workflows.
Cargo
Cargo is GTM infrastructure: declare your go-to-market as TypeScript in git and let coding agents deploy, retry, and trace every play.
Featured Head-to-Head Comparisons
Heym vs Spider Cloud
Choose Heym if you need a self-hosted, visual platform to build multi-agent workflows with RAG and human oversight. Choose Spider Cloud if your priority is fast, cheap, and reliable web data extraction for AI agents—especially if you need Browser AI commands or 1,000+ ready-made scrapers. They solve different problems: Heym orchestrates AI actions, Spider Cloud feeds AI with web data.
Heym vs Presto Voice
For QSR chains seeking proven drive-thru AI that boosts revenue, Presto Voice is the clear choice with measurable ROI and major brand partnerships like Dairy Queen. For developers needing a flexible, self-hosted AI workflow platform with full control over agents, RAG, and data privacy, Heym is unmatched — especially given its rapid iteration and open-source model. Choose based on your domain: physical restaurant automation (Presto) vs. software workflow automation (Heym).
Heym vs Temporal Ai
If you need to orchestrate mission-critical, long-running microservices or AI agents with guaranteed durability and state recovery, Temporal is the battle-tested choice (used by OpenAI, Lovable). If you want a visual, NoCode-friendly AI workflow builder with built-in RAG and MCP, and prefer self-hosting, Heym offers a more accessible, modern alternative. Choose Temporal for reliability at scale; choose Heym for rapid development of AI pipelines.
Alternatives to Heym
View all
Frequently Asked Questions
Used Heym? Help shape our editorial sentiment research.