Forum Backup
Recursively walks every section, thread and page of a vBulletin board, downloading posts, avatars and attachments while stripping session tokens to avoid loops and duplicates.
Arcane Forum Archiver is a recursive crawling and mirroring engine that walks an entire vBulletin board — every thread, every post, every avatar and attachment — and binds it into a permanent, self-contained offline archive. No database, no server, no dependency on a site that might vanish tomorrow.
Each tab in the Archiver's interface performs one distinct rite over your forum's data — from first crawl to final, unified tome.
Recursively walks every section, thread and page of a vBulletin board, downloading posts, avatars and attachments while stripping session tokens to avoid loops and duplicates.
Transforms raw archived pages into clean, navigable local HTML — reconstructing thread coherence without a single external dependency.
The full pipeline in one pass: crawl, mirror media, and render to HTML in a single unattended run.
Recovers structured data — author, timestamp, content, hierarchy — from previously archived pages, and re-exports it as clean JSON or CSV.
A general-purpose mirroring worker that normalizes URLs and rewrites every asset reference into a self-contained local copy of a site.
Fuses hundreds of separately archived pages into one unified, cross-linked static site, with assets embedded and a generated navigation tree.
Binds archived documents together into a single merged PDF tome — the final, printable form of a preserved forum.
The engine room behind the ritual: concurrency, deduplication and memory discipline for boards with tens of thousands of posts.
CSS-selector-driven scanning tracks thread, post and page identifiers to know exactly how deep to go — and when to stop.
Hashing sets track every visited and pending endpoint, so nothing is ever fetched or written twice.
Every remote src and href is rewritten to a local path, with MIME validation to keep images and attachments intact.
Binary persistence runs off the main thread, so ingesting thousands of files never freezes the interface.
Crawling and downloading run on dedicated daemon threads, keeping the UI responsive during massive scans.
Cyclic garbage collection and buffer flushing keep long runs — 10,000+ posts and beyond — stable from start to finish.
The log pane narrates every step of the crawl in real time — threads discovered, media mirrored, duplicates skipped — so you always know exactly where the ritual stands.
Prebuilt binaries for Windows and Linux, or the full Python source for those who prefer to read every rune themselves.
Requires Python 3.13 or 3.14 when building from source, with
tkinter, requests and bs4 available. Prebuilt binaries
bundle everything needed to run standalone.