FLUIDFLUID
  • Introduction
  • Quickstart
  • Why FLUID
  • FAQ
  • What FLUID Is
  • Core Principles
  • Agentic-Native Layer
  • FLUID vs ODCS / ODPS
  • Anatomy
  • Cheatsheet
  • Full Specification
  • Versions
  • JSON Schema 0.7.5 ↗
  • Reference (HTML) ↗
Examples
How-to
What's New
Deck
GitHub
GitHub
  • Introduction
  • Quickstart
  • Why FLUID
  • FAQ
  • What FLUID Is
  • Core Principles
  • Agentic-Native Layer
  • FLUID vs ODCS / ODPS
  • Anatomy
  • Cheatsheet
  • Full Specification
  • Versions
  • JSON Schema 0.7.5 ↗
  • Reference (HTML) ↗
Examples
How-to
What's New
Deck
GitHub
GitHub
  • Concepts

    • What FLUID Is (and Is Not)
    • Core Principles
    • The Looming Crisis of Context
    • The Agentic-Native Layer
    • FLUID vs ODCS / ODPS
    • The Reference Compiler — forge-cli

The Looming Crisis of Context

The "modern data stack" — a disaggregated ecosystem of best-in-class tools — has enabled rapid progress, but is held together by fragile scripts, proprietary configs, and tribal knowledge. This complexity, manageable by humans, becomes a liability in the Agentic Revolution.

Agentic AI — capable of complex reasoning and autonomous tool use — will soon be the primary consumer of enterprise data. Its potential, however, is capped by the quality and reliability of the data it can access.

Key questions

  • How can an agent trust the data it consumes?
  • How does an agent discover the correct data product?
  • How can we govern and audit thousands of autonomous agents accessing sensitive data?

The current landscape, built on disconnected pipelines, offers no scalable answers. Deploying agents atop this foundation is like building a skyscraper on sand.

What's needed is a paradigm shift: from data as pipeline output to data as a product with a contract.

FLUID is that foundational, declarative protocol.


Why FLUID is indispensable in an MCP world

Can't a smart AI just "get the data"? Why bother with data products?

No matter how advanced, an AI agent cannot operate on data it does not understand or trust. Connecting to raw databases is a liability, not an asset. FLUID closes three critical gaps:

GapThe problemHow FLUID closes it
Semantic gapWithout a contract, data is just bits.FLUID's contract and semantics provide essential context — schema, descriptions, business ontology links.
Trust gapHow does an agent know data is correct or fresh?FLUID's quality and SLA blocks provide enforceable guarantees.
Governance gapHow do we control and audit agent access?FLUID's accessPolicy and dynamic policies create a programmatic access-control layer.

Conclusion: AI cannot "just get the data." FLUID provides the machine-readable contracts and policies that transform raw data into safe, trustworthy, and understandable data products.


Where to go next

  • The Agentic-Native Layer — the four agent failure modes FLUID answers deterministically.
  • What FLUID Is (and Is Not) — the declarative protocol and the F.L.U.I.D philosophy.
Edit this page on GitHub
Last Updated: 5/29/26, 5:26 PM
Contributors: fas89, Claude Opus 4.8
Prev
Core Principles
Next
The Agentic-Native Layer