Complete Pages, Preserved Quietly
Arivu now combines direct and isolated rendered capture, keeps article images locally, and presents useful preserved copies without opening a browser tab.
Saving a web page should not depend on whether its useful content happened to be present in the first HTML response.
Arivu now has a complete background capture path for pages that need rendering. Safe direct extraction and an isolated browser service can work at the same time. Direct evidence becomes readable as soon as it is ready, while rendered evidence can replace it only when the rendered result is genuinely stronger.
Strongest Evidence Wins
Web capture is not a simple success-or-failure event. A direct response may contain a complete article. It may contain only a shell that expects JavaScript. A browser may render the real page, or it may encounter a challenge, timeout, or partial result.
Arivu treats those results as evidence candidates rather than accepting whichever one finishes last. Source-native and complete direct evidence remain authoritative when they are stronger. A complete rendered reader can upgrade weak direct extraction. A challenge page, empty result, or failed browser run cannot replace the last good copy.
This also makes reprocessing safer. Existing readable content stays available while a new attempt runs. If an extractor improvement produces equally strong fresh evidence, Arivu can activate it instead of retaining an older copy solely because the quality score tied.
What this means for you: difficult pages can become more complete, while a bad retry does not erase the useful reader you already had.
Article Images Stay with the Reader
Reader content is more than paragraphs. Diagrams, charts, photographs, and figures may carry the central evidence. Arivu now downloads the images referenced by the selected reader, stores them in owner-scoped local storage, and rewrites the archived article to use those copies.
This applies to both rendered and ordinary direct captures. A core-only installation can still retain article figures through the safe direct-fetch path. Complete browser capture additionally resolves the images used by the rendered reader instead of relying on whether they happened to load inside the visible viewport.
Safe MathML remains intact too, so supported equations, matrices, fractions, and superscripts do not collapse into plain text during sanitization.
Preserved Copies Have Different Jobs
The reader now presents preserved copies in one responsive panel:
- Screenshot records visible layout.
- PDF records a paginated rendered document when enabled.
- Source response retains what direct retrieval returned.
- Offline HTML keeps a self-contained downloadable page.
Offline HTML is not executed as a live website inside Arivu. Its preview is transformed into an inert, scriptless sandbox, and oversized pages remain downloadable without being opened in the preview.
These formats complement the readable article. They do not make every capture complete, grant access to authenticated material, or turn a publisher’s page into something you are free to redistribute.
The Browser Runtime Stays Out of Your Browser
Complete capture runs headlessly in a separate native service. It does not open a visible tab, read your browser profile, or receive your browser cookies. Its page requests still pass through Arivu’s bounded safe-fetching controls.
Fresh installer runs on supported Linux hosts enable the service by default after showing the plan. Screenshot and offline HTML preservation start enabled. PDF is disabled by default and requires separate operator configuration because of its additional storage cost. The installer downloads and verifies the matching runtime, checks it before activation, and manages upgrades, rollback, disabling, and removal.
You do not need to install Docker, npm, Node, Chromium, or Monolith yourself. If you prefer the smallest runtime, pass the core-only option during installation or change the choice later with reconfiguration. Direct capture remains available, and disabling the browser service does not delete content or artifacts already stored.
The core Arivu application is still one Go binary with SQLite. Complete capture is an installer-managed addition with a clear boundary, not a new collection of services for you to assemble.
Back Up the Evidence Boundary
Local reader images and preserved copies live beside SQLite in the asset store. Installer backups keep the database and assets together and write a manifest that restore verifies. Portable account-level JSON remains a different tool and does not replace a host backup when retained files matter.
Direct capture remains available independently. Disabling complete capture preserves existing reader content and saved artifacts, and a failed browser attempt does not discard successful evidence. The change is that a saved page can now retain more of what made it useful, quietly, without making browser rendering a prerequisite for the rest of Arivu.
Read Preservation and Sharing for current formats, installer choices, public snapshot boundaries, and backup implications.