Service Guides

• Getting Started — Installation, prerequisites, initial config, and first run.
• Configuration Reference — Complete config schema: watch paths, embedding and rate limiting, vector store and payload indexes, API key, state directory, sources, inference rules, search and hybrid search, reindex, VCS, logging, and shutdown.
• Architecture — Data flow pipeline, component diagram, three-sources-of-truth model (filesystem, enrichment store, Qdrant), processing stages, queues, and shutdown.
• Inference Rules — Rule matching, schema merging, set expressions, JsonMap transforms, Handlebars templates and render, reserved field names, runtime rules, and renderAs.
• Source Documents — Indexing records from other systems by callback: sources config, the callback contract, designated rules, notify and coalescing, reindex, and the SDK.
• API Reference — HTTP endpoint reference for search, documents, config, reindex, scan, metadata, authentication, rate limiting, and diagnostics.
• CLI Reference — All jeeves-watcher commands: core standard commands and domain-specific commands.
• Deployment — Running as a service (NSSM/systemd/launchd), production config, state directory and backups, log management, and updates.
• Version Control (VCS) — Git-backed content versioning, commit batching, squash retention, remote push, and reversion.
• Changelog — Release history for @karmaniverous/jeeves-watcher.