Adopt a single semantic model specification so KPI definitions stay consistent across BI tools and AI agents.
Convert existing semantic models between Ossie and formats like dbt, GoodData or Salesforce using the reference converters.
Validate a semantic model against the Ossie schema before rolling it out across your data stack.
Study the example TPC-DS semantic model as a template for defining your own metrics.
| apache/ossie | beenuar/aisoc | wenyuchiou/awesome-agentic-ai-zh | |
|---|---|---|---|
| Stars | 1,489 | 1,479 | 1,504 |
| Language | Python | Python | Python |
| Last pushed | — | 2026-06-30 | — |
| Maintenance | — | Active | — |
| Setup difficulty | easy | hard | easy |
| Complexity | 2/5 | 4/5 | 1/5 |
| Audience | data | ops devops | vibe coder |
Figures from each repo's GitHub metadata at analysis time.
This is a specification and schema, not a running service, so there is nothing to deploy.
Apache Ossie, previously called Open Semantic Interchange, is an open source effort to create one shared way of describing what data means across the many different tools used in analytics, AI and business intelligence. The problem it addresses is that the same business metric, like a KPI, often gets defined differently in each tool a company uses, which forces teams to spend real time reconciling those different definitions by hand, and can lead AI agents to produce unreliable answers because they are working from inconsistent underlying business logic. The project's answer is a single specification, written in JSON and YAML, that any tool can read and write. The goal is a vendor neutral single source of truth for semantic metadata, so a data field's meaning and value stay the same no matter which tool or AI agent is using it, removing the inconsistencies that come from each platform maintaining its own separate definitions. The repository is organized into a few clear parts. The core-spec folder holds the actual specification document along with its machine readable schema files. A converters folder contains reference code that translates between the Ossie format and other semantic formats used by tools like dbt, GoodData, Polaris and Salesforce. An examples folder includes sample semantic models, including a full model based on the standard TPC-DS benchmark dataset. There is also a validation folder with tooling to check whether a semantic model matches the Ossie schema, and a docs folder with general project documentation. As an Apache incubating project, Ossie is community driven. The README points people who want to get involved toward a CONTRIBUTING guide for proposing specification changes, a ROADMAP document describing current working groups and planned enhancements, and both GitHub Discussions and a Slack community for ongoing conversation. The project is licensed under the Apache License, Version 2.0.
An Apache incubating specification that standardizes how KPI and metric definitions are shared across BI and AI tools.
Mainly Python. The stack also includes JSON, YAML, Python.
Free to use, modify and redistribute for any purpose, including commercial use, as long as you keep the license notice.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly data.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.