How to import your Pocket or Omnivore saves

Most people who saved articles to Pocket or Omnivore did so with good intentions. They spotted something worth reading, tapped a button, and moved on. Over months or years, those saves accumulated into a collection that represents a genuine map of interests, research threads, and half-finished thinking. The trouble is that these tools kept everything in a flat list. Finding something later meant scrolling, guessing at keywords, or giving up entirely.

If you are moving away from Pocket or Omnivore, the question is not just where to put your saved articles. It is whether the new home can do something useful with them. Fabric imports your saves and then connects them to the rest of your knowledge, making the entire collection searchable by meaning and queryable through AI.

Why people leave Pocket and Omnivore

Pocket has been around since 2007. It works well enough as a read-later queue, but its search is limited to titles and tags, and it does not connect your saves to anything else you know or store. Omnivore, which positioned itself as an open-source alternative, shut down its hosted service in late 2024 after being acquired by ElevenLabs. Users were left to export their data before the lights went off.

Both tools share the same fundamental limitation. They treat saved content as isolated items in a list. You can tag them, you can archive them, but the tools never help you understand what you have collected or how it relates to your wider work. A thousand saved articles sitting in a queue are not a knowledge base. They are a backlog.

Fabric takes a different approach. When you import your saves, every article enters a memory engine that maps relationships between your imported content and everything else in your account. An article you saved about climate policy three years ago gets linked to the research paper you clipped last week and the notes you wrote yesterday. The connections form without manual effort.

How the import works

The process depends on which service you are coming from. For Pocket, you export your data through Pocket's settings, which produces an HTML file of your saved links. For Omnivore, you would have exported before the service closed, producing a set of files including your articles, highlights, and labels.

Once you have your export file, you upload it to Fabric. The import process pulls in your saved URLs, any tags or labels you applied, and the content of the articles themselves where available. Fabric stores everything in your cloud drive, so nothing sits on your local device. All content is encrypted with AES-256 and the platform holds CASA Tier 2 certification, which means your reading history is treated with the same care as sensitive business data.

After the import completes, Fabric's indexing runs across your collection. This is where things diverge from what Pocket or Omnivore offered. Rather than just recording that you saved an article, Fabric processes the content so that semantic search can find it by meaning, not just by title or tag. If you search for "arguments against remote work mandates," you will find articles that discuss that topic even if those exact words never appear in the text.

What happens to your tags and folders

Any tags you created in Pocket or labels from Omnivore carry over during the import. Fabric supports both tags and folders, so you can keep your existing structure or reorganise entirely. Many people find that the tagging systems they built in read-later apps were workarounds for poor search. When search works properly, you need fewer tags.

That said, if you spent time building a careful taxonomy, it will not be discarded. Your tags remain attached to each item and work as filters alongside Fabric's semantic search. You can also create new folders to group content by project, theme, or any other logic that suits your workflow.

Making your saves useful with AI

The most significant change when moving from Pocket or Omnivore to Fabric is what happens after the import. In a traditional read-later app, your saved articles sit and wait for you to read them. In Fabric, the AI assistant can summarise, compare, and answer questions about your entire collection.

Suppose you saved forty articles about machine learning over the past two years. You can ask the assistant to summarise the main themes across those articles, identify where authors disagree, or pull out specific data points. The assistant works across your full library, not just one article at a time. This turns a passive reading list into something you can interrogate.

The AI capabilities extend beyond Fabric's own interface. Through MCP integration, you can access your Fabric content from within ChatGPT or Claude. This means you can be working in your preferred AI tool and pull in context from your saved articles without switching apps. You can also grant agent access that is scoped and time-limited, so an AI workflow can read specific parts of your library without having open-ended access to everything.

Building forward from your import

Importing your Pocket or Omnivore saves gives you a foundation, but the real value comes from continuing to add to it. Fabric's web clipper works across browsers and lets you save articles, pages, and selections as you browse. Quick capture handles saving from mobile devices and other contexts where a full clipper is not practical.

Every new save enters the same system as your imported content. The memory engine updates its map of connections as new items arrive, linking fresh articles to older saves where the topics overlap. Over time, the value of your imported Pocket or Omnivore library increases because each new addition makes the existing collection more findable and more connected.

This is the core difference between a read-later app and a knowledge platform. A read-later app is a queue. Items go in, and with luck, they come out again when you remember to look. A knowledge platform treats every saved item as a node in a growing network. The more you add, the more useful the whole system becomes.

Handling large collections

Some long-time Pocket users have thousands of saved links. Omnivore users who imported from other services before it closed may have similarly large libraries. Fabric handles large imports without requiring you to trim your collection first. The cloud drive has no practical storage ceiling for this kind of content, and the indexing process runs in the background without blocking your use of the platform.

If your collection includes links to pages that are no longer available online, Fabric preserves whatever content was captured at the time of your original save. For Pocket exports, this depends on whether Pocket had cached the full article text. For Omnivore exports, full article content is typically included in the export files. Either way, you will not lose content that was already captured just because the original URL has gone dark.

A note on Omnivore specifically

Since Omnivore's hosted service has closed, users who did not export their data before the shutdown may have limited options for recovery. If you still have access to your Omnivore export files, Fabric can process them. If you were using a self-hosted Omnivore instance, you can export from that and bring the data into Fabric the same way.

For those who missed the export window, it is worth checking whether Omnivore sent a final data export by email or made archives available after the closure. The situation varies by user, and Fabric cannot recover data that was lost with the service. But for anyone holding an export file and wondering where to put it, the import path into Fabric is open.

Frequently asked questions

Can I import from both Pocket and Omnivore into the same Fabric account?

Yes. You can run multiple imports into the same account. Content from both services will be indexed together, and the memory engine will find connections across both collections as well as any other content in your account.

Will my Pocket tags transfer to Fabric?

Tags from your Pocket export carry over and appear as tags within Fabric. You can keep them as they are, rename them, or reorganise your content into folders after the import. Nothing is lost in the transfer.

What format does my Omnivore export need to be in?

Omnivore provided exports in a structured format that includes article content, highlights, and labels. If you have the export archive that Omnivore generated, Fabric can process it. The key is that the export files are intact and have not been modified since download.

How long does a large import take?

The import itself completes relatively quickly, even for collections of several thousand items. The indexing that follows, where Fabric processes content for semantic search and maps connections, runs in the background and may take longer for very large libraries. You can use Fabric normally while indexing continues.

Is my imported content private?

All content in Fabric is encrypted with AES-256, and the platform is CASA Tier 2 certified. Your imported articles are stored in your personal cloud drive and are not visible to other users. If you use AI features, the assistant accesses your content within Fabric's secure environment.

Can I search my imported articles by topic rather than keyword?

Yes. Fabric uses semantic search, which understands the meaning of your query rather than just matching exact words. If you search for a concept, you will find articles that discuss that concept even if they use different terminology.

What happens if a saved URL is no longer live?

Fabric preserves the article content that was captured in your Pocket or Omnivore export. If the original page has been taken down or moved, the saved content remains accessible in your Fabric library. However, if Pocket or Omnivore never cached the full text of a particular save, only the URL and title will be available.

Can I access my imported saves from ChatGPT or Claude?

Yes. Through Fabric's MCP integration, you can connect your Fabric library to external AI tools. This lets you reference your saved articles and imported content from within ChatGPT, Claude, or other compatible AI assistants without leaving those tools.

Do I need to keep Pocket installed after importing?

No. Once your data is in Fabric, the import is complete and self-contained. You can delete your Pocket account or simply stop using it. For ongoing article saving, Fabric's web clipper and quick capture handle the same workflow.

How does Fabric differ from just using another read-later app?

Read-later apps store links and, at best, let you search titles and tags. Fabric stores content in a knowledge platform where every item is semantically searchable, connected to your other content through the memory engine, and queryable by AI. Your saved articles become part of an active, growing knowledge base rather than a static list. You can also surface your imported content through AI agents with scoped, time-limited access, which no read-later app offers.

Related pages

• How to replace Pocket

• How to save everything you find online in one place

• How to import your browser bookmarks