Skip to content

Tortoise-AI/uds

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Universal Dashboard Specification (UDS)

Version License Status

A vendor-neutral, declarative format for AI-native analytical dashboards


What is UDS?

Modern organizations face a fundamental problem with their analytical dashboards: they're locked into proprietary platforms, impossible to version control, and can't be generated by AI at scale. Each vendor uses its own format, dashboards are defined through point-and-click interfaces, and migrating between platforms requires complete rebuilds.

The Universal Dashboard Specification (UDS) solves this by providing a vendor-neutral, declarative format for describing analytical dashboards. Think of it as "the Kubernetes of dashboards" — a specification layer that sits above any rendering platform.

UDS enables:

  • Semantic-first design: Dashboards reference metrics and dimensions from your existing semantic layer (Cube, dbt Semantic Layer, Looker, etc.) rather than hard-coding SQL
  • Intent-driven declaration: Every dashboard explicitly declares its purpose, the decisions it enables, and who it serves
  • AI-powered generation: Built from the ground up for AI assistants to generate dashboards with constraints, confidence scores, and human review workflows
  • True portability: Write once, compile to Grafana, Metabase, or custom renderers
  • Version control native: Dashboards are YAML files that live in git alongside your application code
  • Persona awareness: Dashboards adapt their complexity, metrics, and visualizations to executives, managers, analysts, operators, or external users

UDS is not a rendering engine. It's a specification that owns the abstraction layer while letting you choose your rendering platform. This creates a sustainable ecosystem where tooling vendors can build parsers, renderers, and AI agents without platform lock-in.

Key Features

  • Semantic-first: References metrics/dimensions from semantic layer (Cube, dbt, Looker, etc.) instead of raw SQL
  • Intent-driven: Every dashboard declares its purpose and the decisions it enables
  • Persona-aware: Adapts to executive, manager, analyst, operator, and external user personas
  • AI-native: Designed for generation with constraints, confidence scores, and human review
  • Portable: Compile to Grafana, Metabase, Looker, or custom renderers
  • Accessible: WCAG 2.1 AA compliant by design with screen reader support
  • Extensible: Layered conformance (Core/Standard/Complete) allows gradual adoption
  • Version control friendly: Plain YAML that works with git, CI/CD, and code review

Quick Example

Here's a minimal UDS dashboard that shows daily revenue:

uds: "0.1.0"

dashboard:
  title: "Sales Overview"
  intent: "Monitor daily sales performance and identify revenue trends"
  persona:
    type: manager
    customization_level: moderate

  semantic_sources:
    - id: sales
      type: cube
      endpoint: "https://cube.example.com/cubejs-api/v1"
      auth:
        type: bearer
        token_env: "CUBE_API_TOKEN"

  layout:
    type: grid
    columns: 12
    row_height: 60

  panels:
    - id: revenue
      type: kpi
      title: "Total Revenue"
      position:
        row: 0
        column: 0
        width: 4
        height: 2
      data:
        source: sales
        metrics:
          - ref: "sales.total_revenue"
        time_range:
          type: relative
          value: "today"
      formatting:
        number_format:
          style: currency
          currency: USD

This dashboard:

  • Declares its intent (monitoring sales performance)
  • Specifies it's for a manager persona
  • References metrics from a Cube semantic layer
  • Uses a responsive grid layout
  • Includes a single KPI panel with currency formatting

Documentation

  • Specification: Complete UDS v0.1.0 specification
  • Examples: See the examples/ directory for sample dashboards
  • JSON Schema: Coming in Phase 3
  • Tooling: Reference parser and renderer coming in Phases 4-5

Project Status

Current Version: v0.1.0 (Draft)

Roadmap

  • Phase 1: ✅ Specification complete (v0.1.0)
  • Phase 2: Examples and sample dashboards (in progress)
  • Phase 3: JSON Schema for validation
  • Phase 4: Reference parser (TypeScript)
  • Phase 5: Reference renderer (React)
  • Phase 6: Conformance test suite
  • Phase 7: Documentation site
  • Phase 8: Public launch and community building

Related Projects

Future ecosystem projects (coming soon):

  • uds-parser: Reference TypeScript parser for UDS documents
  • uds-renderer: React-based reference renderer
  • uds-validator: JSON Schema validator and conformance checker
  • uds-ai-agent: AI assistant for generating UDS dashboards
  • uds-compiler-grafana: Grafana JSON compiler
  • uds-compiler-metabase: Metabase API compiler

This specification accompanies From Policy to Practice: An Open Framework for AI-Ready Project Delivery (Newman, 2026). DOI: https://doi.org/10.5281/zenodo.18711384

Use Cases

UDS is designed for:

  • AI-powered dashboard generation: LLMs generate dashboards from natural language with built-in constraints and review
  • Platform migration: Move dashboards between Grafana, Metabase, Looker, or custom platforms
  • Version control and CI/CD: Store dashboards in git, review changes, and deploy via pipelines
  • Multi-tenant analytics: Generate personalized dashboards for different user personas
  • Semantic layer adoption: Decouple dashboards from raw SQL and reference your semantic layer
  • Compliance and governance: Enforce data access policies, accessibility standards, and documentation requirements

Contributing

UDS is in early development. We welcome contributions from:

  • Dashboard platform vendors
  • Semantic layer providers
  • AI/ML companies building generation tools
  • Data practitioners and dashboard designers
  • Open source enthusiasts

See CONTRIBUTORS.md for the list of contributors.

Note: A Contributor License Agreement (CLA) will be required for all contributions. Details will be published as the project matures.

Citation

If you use this specification in your research or work, please cite:

Newman, A. (2026) From Policy to Practice: An Open Framework for AI-Ready Project Delivery.
London: Tortoise AI. DOI: https://doi.org/10.5281/zenodo.18711384

License

Copyright 2026 Tortoise AI Ltd

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at:

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the LICENSE file for the specific language governing permissions and limitations under the License.


Built by Tortoise AI Ltd | GitHub | Website

About

No description, website, or topics provided.

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors