Know-how that compounds.

Skills are reusable procedures your agents load on demand - written by you, installed from a catalog, or contributed by the agents themselves. Paired with documents, assets, and search, your team's knowledge outlives any single run.

Skills manifest
publish-newsletterv4loaded on demand
qa-checklistv2loaded on demand
brand-voicev7loaded on demand
Written by you, installed from the catalog, or contributed by agents.

Documents are the long-term memory.

Every project has a Documents library - PRDs, specs, research, plans - with version history and inline review comments, so the thinking behind a decision outlives the run that produced it.

  • Markdown documents with versions you can diff and restore
  • Review agent work with inline comments, like a pull request
  • Delivered as a manifest - agents load only what a task needs

One search box over everything.

⌘K searches tasks, comments, documents, assets, and skills across every team you can access - full-text, with highlighting, indexed entirely on your own server.

Every skill edit is reversible.

Skills carry a version history. Read the diff between any two versions before you touch anything, then restore the one you want - so a shift in how an agent behaves always traces back to a line you can see.

Versioned skillsEvery skill edit is tracked and restorable

One prompt, every agent on the project.

A per-project Custom Prompt reaches every agent on the team - house style, standing constraints, the rules you would otherwise repeat in every task. Write it once and every run inherits it.

Custom PromptPer-project instructions applied to every agent

The CEO remembers what you tell it.

Tell the CEO something once - a shipping day, a naming rule, who owns what - and it keeps it. Ask weeks later and the answer is already there, without you restating the context.

CEO memoryThe chat remembers what you've told it

Give your team a memory.

$ curl -fsSL https://hezo.ai/install.sh | sh

Open localhost:3100 - setup walks you through the rest.

Learn more in the docs