kotobase
Menu

Kotobase engineering blog

Technical decisions, reproducible experiments, integration guides, and release evidence—not generic graph-database SEO.

Latest

Why an Obsidian graph should not own your Markdown

A local-first design for conflict-safe two-way sync, queryable [[wikilinks]], and CID-reproducible AI context.

See the Obsidian workflow · Read the implementation guide (opens on GitHub)

One tenant graph, multiple query protocols

How Cypher, Gremlin, GraphQL, SPARQL, and Datomic queries share identity, capability discovery, and audit boundaries.

Read the protocol evidence (opens on GitHub)

A signed Worker is only useful if you deploy those exact bytes

Clean rebuilds, generated-source drift checks, SBOM/SLSA attestations, and exact-artifact deployment.

Read the supply-chain guide (opens on GitHub)

Enterprise readiness without compliance theatre

A living gap register that separates implemented controls, contractual commitments, and work that still requires independent evidence.

Read the current register (opens on GitHub)

Editorial contract

Every post must link to runnable code, an ADR, measured evidence, or a clearly labeled proposal. Marketing claims do not become facts by appearing here.