The whole ecosystem, in one place
16 on npm and 2 on PyPI, grouped by the job they do. What each one is, where to install it, and where to read more if there is more to read.
9 of them have a page here that goes deeper. The rest link straight to their registry and repository.
The execution layer
A model asks for an action by name; the runtime decides whether it may, performs it, and writes down what happened. Everything else on this page is built on these.
@openverb/runtimenpmThe execution runtime. Resolves a named verb, runs it, and records the result.
@openverb/policynpmThe policy engine. Decides whether a requested action is allowed before it runs.
The document format
A .ov file is a sequence of actions a runtime can execute — which is what makes a document portable between hosts rather than tied to the tool that wrote it.
Verb libraries
A runtime on its own knows no verbs. These are the vocabularies: what a model is allowed to ask for in a given domain.
@openverb/stdlibnpmThe standard library — theme, navigation, search, modals, forms, sessions and focus. The starting vocabulary for any interface.
@openverb/idenpmVerbs for AI-native coding environments: files, the editor, git, the terminal and preview.
@openverb/gisnpmVerbs for geographic work — layers, cameras, selection and measurement.
@openverb/crmnpmA developer-native CRM template, protocol-compliant, with AI-driven communication workflows.
Geographic storytelling
One profile of the format, and the renderer that draws it. GeoStory is what a document says; Motion is what a frame looks like.
Tooling
Things built on the execution layer that are useful on their own, and double as evidence that the layer carries real work.
Python
The same ideas where the agents actually live. Published to PyPI rather than npm, so they are counted separately.
Download figures are last-week totals, read live from the npm and PyPI APIs when this page loads — the same source as the weekly stats. They count installs, which is not the same as use: a package can be pulled by a build a hundred times and never run by anyone. Shown here so the list reads as something alive, not as a measure of adoption.