docs: announce planned Apache Ossie support

docs.json

@@ -184,6 +184,7 @@
             "pages": [
               "guides/lightdash-semantic-layer",
               "guides/lightdash-yaml",
+              "guides/apache-ossie",
               {
                 "group": "Developer quickstart",
                 "icon": "laptop-code",

guides/apache-ossie.mdx

@@ -0,0 +1,61 @@
+---
+title: "Apache Ossie support"
+icon: "handshake"
+description: "Lightdash will support Apache Ossie (Incubating), the open specification for semantic layer and ontology interchange."
+---
+
+<Note>
+  Apache Ossie is currently incubating at the Apache Software Foundation and its spec is pre-1.0. Support in Lightdash is planned — this page describes what we intend to support and how it relates to the Lightdash semantic layer today. Timelines will follow the spec's stabilization.
+</Note>
+
+## What is Apache Ossie?
+
+[Apache Ossie (Incubating)](https://ossie.apache.org) is an open, vendor-neutral specification for exchanging semantic models between tools. It defines a shared, machine-readable format for describing datasets, relationships, fields (dimensions), and metrics — plus the business context around them — so that any tool in a data stack can produce or consume semantic definitions without losing meaning.
+
+The project sits under the Apache Software Foundation, is developed in the open on GitHub, and is backed by a coalition of data and analytics vendors. Its goal is to make a business concept like "Monthly Active Users" resolvable consistently across warehouses, BI tools, and AI agents, regardless of where it was originally defined.
+
+## How Ossie relates to the Lightdash semantic layer
+
+The [Lightdash semantic layer](/guides/lightdash-semantic-layer) already models the same core building blocks Ossie standardizes:
+
+| Ossie concept       | Lightdash equivalent                                                                     |
+| ------------------- | ---------------------------------------------------------------------------------------- |
+| Semantic model      | A Lightdash project's models and their [YAML](/guides/lightdash-yaml) / dbt definitions  |
+| Dataset             | A [table / model](/get-started/develop-in-lightdash/adding-tables-to-lightdash)          |
+| Relationship        | [Joins](/references/joins) between models                                                |
+| Field (dimension)   | [Dimensions](/references/dimensions)                                                     |
+| Metric              | [Metrics](/references/metrics)                                                           |
+| AI context          | Descriptions, [AI hints, and verified content](/references/writing-descriptions)         |
+| Custom extensions   | Model-level `meta` / configuration extensions                                            |
+
+The overlap is close enough that Lightdash semantic models can be expressed in Ossie's format, and Ossie models authored elsewhere can be mapped onto Lightdash's model. Where the two differ today, Ossie is a portable interchange format while Lightdash is a full query engine, catalog, and UI on top of the semantic layer.
+
+## What Lightdash will support
+
+We plan to support Apache Ossie as a first-class interchange format for the Lightdash semantic layer. The intent is that you should be able to:
+
+- **Import** an Ossie semantic model into a Lightdash project so its datasets, relationships, dimensions, and metrics show up as queryable content, with descriptions and AI context preserved.
+- **Export** a Lightdash project (whether defined via dbt or [Lightdash YAML](/guides/lightdash-yaml)) as an Ossie semantic model, so other Ossie-compatible tools and AI agents can consume the same definitions.
+- **Round-trip** models between Lightdash and other Ossie-compatible tools without losing meaning, using Ossie's `custom_extensions` mechanism to carry Lightdash-specific configuration that isn't part of the core spec.
+
+Because Ossie is still evolving, our implementation will track the spec as it stabilizes rather than pinning to a single early draft.
+
+## Following Ossie's roadmap
+
+The Ossie community is actively working on a set of areas that will directly benefit Lightdash users as they land:
+
+- A more expressive core model — richer metric semantics, explicit grain and entity modeling, more capable relationship definitions, and reusable semantic filters.
+- A standardized semantic query language and reference engine, so consumers (BI tools, AI agents, APIs) can query any Ossie-compatible semantic layer through a consistent interface.
+- Catalog integration and semantic services, making Ossie models discoverable and governable at the platform layer (for example, alongside Apache Polaris).
+- An ontology layer that describes business concepts independently of physical tables, enabling conceptual — not just structural — interoperability across models.
+- Richer field semantics (units, currencies, PII classification, display metadata) and better tooling (validators, converters) around the spec.
+
+We'll evolve Lightdash's Ossie support in step with these directions, prioritizing the pieces that unlock better interoperability, AI grounding, and governance for Lightdash customers.
+
+## Where to learn more
+
+- Apache Ossie project site: [ossie.apache.org](https://ossie.apache.org)
+- Source and spec: [github.com/apache/ossie](https://github.com/apache/ossie)
+- Lightdash semantic layer: [Lightdash semantic layer](/guides/lightdash-semantic-layer)
+
+If you're evaluating Ossie for your stack and want to talk through how it will fit with Lightdash, [get in touch](/contact/contact-info).