Skip to main content

Use Cases

ALX Protocol defines deterministic identity and verifiable relationships for data produced by intelligent systems. AI agents and applications represent prompts, model responses, tool results, decisions, workflow states, agent handoffs, source documents, datasets, analyses, reviews, and other data objects as portable ALX Blocks.

Each Block can declare other Blocks as direct parents through parentHashes. Declared parent relationships form Block Graphs, allowing applications to inspect, validate, and traverse declared lineage independently of the originating system.

Block Structure​

Each Block contains five fields:

  • protocolVersion — identifies the ALX protocol version used to construct and validate the Block.
  • content — JSON-serializable application-defined data.
  • parentHashes — references to directly declared parent Blocks.
  • contentHash — deterministic hash of the canonicalized Block content.
  • blockHash — deterministic hash derived from Block content and normalized parent relationships.

ALX uses deterministic hashing to establish Block identity. Changes to content change contentHash and blockHash. Changes to parentHashes change blockHash. Equivalent protocol inputs produce identical hashes across conforming implementations.

Applications can independently recompute Block hashes and validate Block structure. When parent Blocks are available, applications can also traverse and validate declared relationships. ALX supports local Block creation and validation without a blockchain, token, hosted service, storage provider, or vendor API.

Protocol validation confirms deterministic identity, Block structure, and declared relationships. Validation does not establish factual accuracy, creator identity, ownership, or the completeness of real-world dependencies.

Where ALX Applies​

  • AI & Agent Workflows — AI agents produce retrieved sources, tool results, intermediate outputs, and final answers. Agent frameworks often store supporting context in private logs or application-specific databases, limiting independent verification and reuse across systems.

  • Software Provenance — Software releases are built from source revisions, resolved dependencies, and toolchain configurations. Most build systems record these inputs in job logs, build databases, or registry metadata. Other systems cannot independently verify the inputs that produced a release.

  • Research & Data — Research workflows produce datasets, cleaned tables, analysis results, and published figures. Reconstructing the relationships between these outputs often depends on documentation, file names, and application-specific records.

  • Compliance & Audit Records — Organizations exchange approvals, assessments, and compliance evidence across independent systems. Application-specific identifiers and audit logs complicate verification across organizational boundaries.

Start Building​

Quick Start creates and validates a Block. Compose Blocks into a Graph declares parent relationships and verifies the resulting Graph.