FCP Introduction
FCP is the standard for graphing context into trusted World Models.
Today, our digital world is fragmented. Every person, application, and agent maintains their own isolated view of reality. This separation creates context amnesia—where knowledge is locked in silos, and trust cannot travel.
Fide Context Protocol (FCP) defines the standard for graphing context—capturing not just inputs and outputs, but the decision-making process—into trusted World Models.
It allows both Human and AI Actors to build their own verifiable view of reality—whether private, public, or a mix of both. Instead of a single centralized truth, FCP enables an Fide Context Graph where reality is subjective: what you see depends on who and what you choose to trust.
What is a 'World Model'?
In AI, this term is used to describe three different things:
- Physical Simulators (like LeCun's JEPA): Systems that predict consequences and cause-and-effect.
- Generative Models (like Sora): Media models that understand 3D space, continuity, and object permanence.
- Knowledge Graphs: Structured maps of facts, relationships, and "ground truth."
Here is the bottom line: Regardless of which definition you use, every World Model needs reliable data to function.
FCP is the standard for graphing context—capturing the Actions and Relationships needed to build trusted World Models of any type.
Fide Context Graph
Instead of locking data in silos, FCP sets the standard for building a distributed Fide Context Graph from which trusted World Models are created.
Structure
The graph is built from three core elements:
- Entity Types (entities) — Strictly defined types organized into Active, Inactive, and Protocol entities
- Statement Schema (edges) — The unified pattern with Attributes, Relationships, and Actions connecting entities
- Fide ID Identifiers — Universal, content-addressed identifiers enabling location-independent entity recognition
Lifecycle
Data flows through four stages:
- Drafting — Create unsigned Statements using the standardized schema
- Attesting — Sign statements with cryptographic proof of authorship (creating Attestations)
- Broadcasting — Publish attestations to distributed registries for discovery
- Indexing — Ingest, verify, and materialize attestations into a queryable store
Applications
Three primary use cases:
- Query & Resolve — Retrieve, resolve, and traverse data using entity emergence
- Consensus — Aggregate trust signals to determine ground truth
- Evaluating — Create signals for the graph (Methods, Statements)
Build with FCP
Ready to start building? Use our official libraries to ensure protocol compliance and simplify development.
JavaScript / TypeScript SDK
The production-ready SDK for Node.js and Browser. Handles Fide ID calculation, signing, and attestation creation.
SDK Overview
Learn about our language support, compliance guarantees, and planned SDKs.