Cloud-native, auditable, schema-first entity-graph-relational data store for agentic applications and business workflows.
⚠️ Pre-release. The wire protocol, data model, and SDK surface are not yet frozen. Breaking changes may land in any minor version until v1.0 without a deprecation period. Production use is at the operator's own risk. Pin a specific version and track the CHANGELOG for API-impacting changes. See the Product Vision — Stability Status for the full stability contract.
Status: Early design phase
Axon is the central nervous system for agentic applications. It provides a transactional data store where entities (deeply nested, schema-validated structures) are connected by typed links (directional relationships), with ACID transactions, immutable audit trails, and enforced schemas as guarantees.
Agents and humans share a trustworthy substrate for structured, interconnected state.
| Document | Description |
|---|---|
| Product Vision | Mission, vision, target market |
| PRD | Requirements, data model, transaction model, value propositions |
| Principles | Design principles with testable criteria |
| Technical Requirements | Stateless servers, multi-backend, data shape limits, correctness, ESF schema |
| Competitive Analysis | 18-system comparison across graph, relational, NoSQL, and hybrid databases |
| Feature Specs | FEAT-001 through FEAT-010 |
| Use Case Research | 10 domains: CRM, CDP, AP/AR, ERP, Issue Tracking, and more |
| Schema Format Research | 19 schema formats evaluated (JSON Schema, OWL, SHACL, EdgeDB SDL, etc.) |
| FoundationDB DST Research | Deterministic simulation testing approach for correctness |
| ADR-001: Rust | Implementation language decision |
| ADR-002: Schema Format | Hybrid JSON Schema + Axon vocabulary |
| SPIKE-001: Backing Stores | PostgreSQL, SQLite, FoundationDB, fjall evaluation |
| Test Plan | Correctness invariants, business scenarios, property tests, benchmarks |