archive-crawler

What problem does it solve?

archive-crawler turns scattered personal files (writing, ideas, relationships, and meaningful conversations) into searchable brain pages without you manually sorting through every folder and document.

Core Features & Use Cases

• Universal archive crawling across local mounts, Dropbox/B2-style storage, Gmail takeouts, and mailbox/PST exports to discover “gold” content inside large personal collections.
• Gold filtering with noise skipping so it focuses on user-authored writing, ideas, relationship material, and origin stories while avoiding low-value items like system/config files, spam, and binary-heavy artifacts unless you ask.
• Explicit safety allow-list enforcement that refuses to run unless gbrain.yml defines archive-crawler.scan_paths: to prevent over-scoping sensitive content.
• Manifest-driven triage and review loop that inventories the tree, proposes a priority queue, captures exact user reactions, ingests approved items, and never re-shows already triaged content.

Quick Start

Tell the agent: "crawl my archive and surface the writing worth keeping," after you set archive-crawler.scan_paths: in gbrain.yml to your allowed folders.