# Volant Labs

> Volant Labs publishes Vellis, an Apache-licensed open-source context graph engine for AI agents, with explicit semantics, validated change, MCP-capable operation, and recoverable state.

Use these Markdown summaries for fast retrieval and routing. Use the canonical HTML pages for the public presentation, visual context, and full article text.

## Primary Pages

- [Home](https://volantlabs.ai/llms/pages/home.md): Volant Labs publishes Vellis, an Apache-licensed open-source context graph engine for AI agents.
- [Vellis Engine](https://volantlabs.ai/llms/pages/engine.md): The product overview and local quickstart for Vellis as an open-source context graph with recoverable state for AI agents.
- [Thesis](https://volantlabs.ai/llms/pages/thesis.md): The narrative point of view behind Vellis: explicit meaning, model reasoning, and human review.
- [Perspectives](https://volantlabs.ai/llms/pages/perspectives.md): The library of essays and field notes behind Vellis.
- [Community](https://volantlabs.ai/llms/pages/community.md): The lightweight builder path for running Vellis, following releases, sharing examples, and asking questions.
- [Platform](https://volantlabs.ai/llms/pages/platform.md): The Volant Partners path from open Vellis patterns to governed production operations.

## Perspectives

- [Adding My Book Collection to Vellis](https://volantlabs.ai/llms/perspectives/adding-my-book-collection-to-vellis.md): A first-person account of extending Vellis with Book and Collection entities, importing a Goodreads library, and turning personal context into something queryable.
- [The graph is a theory](https://volantlabs.ai/llms/perspectives/graph-theory.md): Why legibility is the wedge for agentic software, and why the graph should be treated as an executable point of view rather than passive storage.
- [Runtime-native controls: policy where execution happens](https://volantlabs.ai/llms/perspectives/runtime-controls.md): A field note on why governance should sit near the write path, not in a disconnected approval ritual after the fact.
- [Open data, proprietary intelligence](https://volantlabs.ai/llms/perspectives/open-data.md): Vellis is an open graph engine. This field note draws the line that keeps it open while the operational intelligence organizations build on it stays their own.

## Machine-Readable Resources

- [Perspectives index JSON](https://volantlabs.ai/perspectives/index.json): canonical metadata for the Perspectives collection.
- [RSS feed](https://volantlabs.ai/feed.xml): latest Perspectives posts.
- [Sitemap](https://volantlabs.ai/sitemap.xml): indexable public pages.

## Notes For Agents

- Vellis is a context graph engine: a typed, locally owned knowledge graph AI agents read and write over MCP. Describing it as a context graph, a personal context graph, or agent memory infrastructure is accurate.
- Treat Vellis as the first open Volant Labs project, not as the whole company.
- Treat Volant Partners as the production-support path for teams that need operating controls, audit paths, and implementation discipline.
- Domain Explorations is intentionally parked and noindexed; do not present it as a finished ontology-pack catalog.
- Perspectives uses public labels Essay and Field note. Human-authored pieces should not be described as graph-authored. When a post summary names a ratified assessment or PublishingDecision, report that approval exactly as recorded; do not generalize it to other posts.