Skip to content
Omicron

Using Omicron

Import and export

Move posts between Omicron instances, import existing publications, and download portable content archives.

Open Settings → Content (/settings?tab=content) in the web app. Exporting and importing copies content; it does not move an ActivityPub account or its followers, rename a username, or delete anything at the source.

Export your content

  1. Sign in to the source instance and open Settings → Content.
  2. Select Download content and save omicron-content.zip.
  3. Keep the original instance and archive until you have verified the imported posts and images at the destination.

The export contains your locally authored published posts, drafts, and scheduled posts. It preserves their HTML, titles, creation/update dates, tags, language, description, AI-assistance disclosure, and cover-image metadata and credits. It does not export your profile, credentials, sessions, comments, likes, reading lists, followers, subscribers, or payments.

File or directory Contents
omicron.json Version 1 manifest for importing into another Omicron instance
posts/ Standalone HTML files
media/ Local uploaded images referenced by post bodies or covers
wordpress.xml WordPress WXR export
README.txt Format notes and destination limitations

A content archive is not an instance backup. Administrators still need database, upload, and configuration backups.

Import into Omicron

  1. Create or sign in to an account on the destination instance.
  2. Open Settings → Content and choose the export file.
  3. Select Preview import. Review the post titles, source statuses, dates, duplicate markers, image count, and warnings. The preview writes nothing.
  4. Leave Keep previously public posts published unchecked to import drafts. Enable it only after reviewing the publication warnings.
  5. Select Import … posts, then open Manage your posts to review the result.

Supported sources

Source File to choose Notes
Omicron The complete export ZIP Keep omicron.json and bundled images together.
Medium The account export ZIP Reads HTML posts from posts/. All Medium posts become drafts because the export does not reliably identify member-only content.
Substack The publication export ZIP Requires posts.csv and the corresponding posts/<post_id>.html files. Audience-restricted posts remain drafts. Subscriber CSV files are not imported.
WordPress A WXR .xml export Imports posts, not pages or attachment records. Password-protected and non-published posts remain drafts.
RSS An RSS .xml file Imports only the entries and content present in that feed, which may be a recent subset or excerpts. Atom is not supported.

Imported HTML passes through Omicron’s sanitizer. Unsupported scripts, embeds, and styling are removed; review formatting before publishing. Bundled raster images are uploaded to the destination and their references are rewritten. External images remain linked; the importer never fetches their URLs.

Supported image alternatives, passage language/direction, captions, and table header references can survive the content pipeline, but not every source format or destination preserves every relationship. When opening imported HTML in the rich editor would lose detected structure or accessibility information, it stays read-only until you choose Edit simplified content. Opening it or changing only its title does not replace its stored body. Keep the original archive before accepting simplification; review the resulting article before publishing. See content-loss warnings.

Publication and dates

  • Imports default to drafts, even when the source was published.
  • Opting in restores published status only for titled, unrestricted source posts whose creation date is not in the future. These posts go live immediately and may be delivered to followers.
  • Scheduled posts always become drafts; the importer never activates a schedule.
  • Content marked restricted, and all Medium imports, remain drafts even with publication preservation enabled. Review access restrictions before manually publishing them.
  • Valid source creation/update dates are retained on import. An absent or invalid date uses the import time, with a preview warning for the creation date. Publishing an imported draft later follows the normal editor rule: its first publication is dated from that later publication time.
  • Tags are normalized; only the first five are imported. Descriptions are shortened to 150 characters when necessary. The preview warns about truncation.

Duplicates and retries

Source identities are tracked per destination account. A repeated import skips already known posts; it never overwrites an existing post or updates it with a new source revision. Re-exporting imported content keeps that import identity, so a chain of instance-to-instance imports still detects it.

The result reports imported, skipped, and failed posts. Each post and its tags commit together, but the whole archive is not one transaction: successful posts remain if another fails. Fix the reported problem, preview the same file again, and retry. Already imported posts are skipped. Missing/invalid bundled images are rejected before posts are written. Destination upload quotas still apply.

Moving to another platform

Destination Available route
Another Omicron instance Import the complete ZIP using the steps above.
WordPress Import wordpress.xml with its WordPress importer. Image URLs in WXR point to the original instance; keep it online while the destination copies images.
Medium Use Medium’s importer with individual published article URLs, or copy the exported HTML.
Substack Use its importer with a supported website or RSS feed URL, or copy the exported HTML.
Other tools Use the HTML files or the versioned manifest, subject to the destination’s supported formats.

Medium and Substack are not guaranteed to accept the Omicron ZIP directly. Their destination importers determine which URLs, formatting, and metadata they accept. Review the resulting publication before removing the source. No route here transfers paid subscriptions, payments, or social relationships.

Limits

Limit Value
Posts per archive 1,000
Input/compressed export size 50 MiB
Expanded ZIP size 100 MiB
Individual ZIP file 10 MiB
ZIP entries, including directories 5,000
Bundled image 5 MiB; PNG, JPEG, WebP, or GIF
Post HTML before sanitizing 1,000,000 characters
Imported title 1,000 characters
Transfer requests per account 6 per minute, shared by export, preview, and import

The UI labels binary-megabyte limits as MB. Storage quotas can impose a lower effective image allowance. These archive limits are fixed in source; there is no dedicated environment variable to raise them. A larger account export fails instead of silently dropping posts. See Rate limits and the HTTP API.

Found a mistake?Edit this page on GitHub.