Skip to content

Why SEOCHO

Source mirrored from seocho/docs/WHY_SEOCHO.md

SEOCHO is not trying to be the easiest generic memory demo.

It is built for teams that want a graph-native, ontology-governed contract for agent indexing and querying.

One ontology should control the parts of the system that usually drift apart:

  • extraction prompts
  • query planning prompts
  • SHACL-derived validation
  • graph constraints
  • denormalization rules
  • runtime semantic artifacts

In SEOCHO, the same ontology object can be used locally in the SDK and promoted into runtime-safe artifacts instead of maintaining a second schema layer.

Generic AI memory is good at low-friction demos. It is usually weaker at:

  • enforcing a stable graph contract
  • making query behavior inspectable
  • keeping extraction and querying aligned
  • governing schema evolution over time
  • working cleanly with Neo4j or DozerDB as the primary store

SEOCHO is stronger when you care about those tradeoffs.

The ontology is not just reference documentation. It feeds:

  • extraction context
  • query context
  • validation
  • constraint generation
  • runtime artifact generation

SEOCHO assumes graph storage is central, not an afterthought. Local SDK mode and runtime APIs are designed around Bolt-backed graph stores such as Neo4j and DozerDB.

The runtime path is not only prompt text. SEOCHO can derive approved artifacts, typed prompt context, and semantic drafts from the same ontology contract.

SEOCHO is ontology-aware today and is moving further toward deterministic query profiles instead of relying only on free-form text-to-Cypher behavior.

  • internal enterprise knowledge graphs
  • regulated or audited data environments
  • graph-backed agent workflows that need repeatability
  • teams already committed to Neo4j or DozerDB
  • projects that need ontology governance, diff, and migration warnings
  • lowest-friction zero-config memory demos
  • vector-only retrieval abstractions
  • tools that intentionally avoid schema and governance

Choose the surface that matches your job:

SEOCHO is strongest when these three surfaces stay aligned:

  1. GitHub and repo docs define the ontology and runtime contract.
  2. The local SDK uses that ontology directly for indexing and querying.
  3. The runtime consumes approved artifacts built from the same ontology.

That is the product story: one ontology, one graph contract, one runtime path.