Server-side workflows

Run memory pipelines without running backend servers.

Use functions for ingestion hooks, claim extraction, validation workflows, scheduled rebuilds, embedding jobs, and controlled model execution.

Deploy workflow Function examples

export default async function ingest(request) {
  const source = await request.json()
  await maludb.sources.create(source)
  await maludb.claims.extract({ source_id: source.id })
}

Ingestion hooks

Connect email archives, chat systems, repositories, ticket queues, logs, and databases to MaluDB source packages.

Extraction and validation

Run model-assisted extraction, confidence checks, human review queues, and promotion logic near the memory core.

Operational rebuilds

Recreate embeddings, relationship edges, retrieval indexes, and skill packages as policies or models improve.

Built-in observability

Know what each workflow changed.

MaluDB workflow events can be linked to source packages, claims, facts, and audit trails so teams can inspect why memory changed over time.

Logs

Trace source ingestion and extraction outcomes.

Queues

Retry long-running derivation and embedding work.

Audit

Record model, prompt, source, and operator context.