Frequently asked questions
Questions about H1VE
What the method is, how it compares, and whether it fits your team. For questions about a specific tool, see that tool's own docs.
The basics
What H1VE is, and what it is not.
H1VE is a framework for building software with teams of humans and AI agents working as one — with the specifications, gates, and traceability that keep generated code trustworthy. Humans and AI Agents · 1 Team.
No. H1VE is a method — a body of knowledge: a manifesto, principles, roles, a work cycle, artifacts, and metrics. Tools can implement H1VE, but the method itself is something you adopt, not install. This mirrors how Scrum is a method and Jira is a tool.
Because AI fails in ways human-only teams never planned for: it invents fields, renames columns without migrations, produces code that passes a test but corrupts data, and generates faster than anyone can review. Existing methods assume humans write every line. H1VE is built around the specific failure modes of generated code.
One team. Humans and AI agents are not two camps handing work back and forth — they are a single team with a shared cycle, shared artifacts, and clear roles. The "1" is the whole point.
How it compares
Where H1VE sits next to what you already know.
Agile and Scrum organize human collaboration — sprints, standups, backlogs. They are silent on AI. H1VE organizes human + AI collaboration: who approves what the machine produces, how generated code is validated, and how every AI-assisted decision stays auditable. You can run H1VE alongside Agile — they answer different questions.
Code review checks code a human wrote. H1VE's gates are designed for code an agent wrote — which is why validation is split: one role checks function, another checks data integrity, because AI breaks those independently. Plus the AI declaration records what was generated and how much was reviewed, which ordinary review never captures.
Spec-driven development is one piece of H1VE — the principle that no code is written without an approved specification. H1VE adds the rest: the roles, the multi-gate validation, the traceability, and the team structure around that core idea.
Does it fit me
Team size, context, and when not to use it.
No. The method scales down to a solo founder wearing several hats and up to a large organization with strict accountability. On a small team, one person may be both founder and architect — but the responsibilities never merge: a reviewer is never the sole validator of their own work. The control it provides is exactly what enterprises need, and small teams benefit from too.
For a solo throwaway prototype no one will maintain, a one-off script, or a team that doesn't use AI agents at all. H1VE's value comes from governing generated code at some scale of risk. No risk, no need for the gates. Honesty about fit is part of the method.
Speed is not the question — trust is. Teams that adopt AI without governance ship fast and accumulate silent bugs and technical debt. H1VE keeps the speed and adds the judgment, so velocity stops being a liability. The goal is to ship fast and be able to stand behind what you shipped.
The ritual is light: read the context, confirm an approved spec, declare what the AI generated. Much of it happens without leaving the terminal or the AI. What it removes is far costlier — rework from invented decisions, silent data corruption, and "who changed this?" investigations.
In practice
Tools, AI, and how adoption actually works.
No — H1VE is tool-neutral. You can practice it with whatever you have. That said, a reference implementation makes it far easier to operate the cycle and the gates. Reference tools exist; the method does not depend on any single one.
Any capable AI agent. H1VE describes how a team governs AI-generated work — it does not require a particular model or vendor. The agent runs where the developer works; the method governs the flow around it.
Read the manifesto and principles (free, always). Then the playbook for the full method. Use the H1VE Canvas to design your project on one sheet, which gives rise to your foundation documents. Then run your first slice through the cycle. The Get Started guide walks the whole path.
Yes — the method is published in the open. The manifesto, principles, and playbook are there to read and adopt. The standard grows by being used, not by being locked away.
Certification
Credentials, exams, and proving mastery.
Yes — a credential ladder is launching: a shared foundation (HCP), role tracks for Developer, Quality, Data, and Lead, and an expert Architect tier (HCA). They're rolling out in 2026; you can join the waitlist now.
Because it is new. No one holds a recognized credential in multi-agent development yet. Certifying early makes you an authority in the category, not a follower of it — and gives employers a signal that "used AI to code" can't.
No. The exams test the method — judgment under real scenarios, not button-clicking in one product. You can prepare by practicing on any reference implementation.
Pricing is being set. Waitlist members get early-access terms. The anchors — manifesto, principles, playbook — are free to study regardless.
Still have a question the method should answer?
Ask it