Fabric vs Omnivore

Last updated April 2026

Omnivore was a genuinely good read-later app. It shut down in November 2024 after the team joined ElevenLabs, and a lot of people lost a tool they relied on daily. If you're still looking for a new home for your reading and research workflow, here's how Fabric compares to what Omnivore offered, and where it goes further.

Comparison table

What was Omnivore?

Omnivore was a free, open-source read-later app that earned a devoted following. You saved articles, PDFs, and newsletters from the web, and it stripped them down to clean, readable text. It had good highlights, solid full-text search, and the best text-to-speech in the category. The Obsidian integration made it especially popular with the personal knowledge management crowd. In October 2024, the team joined ElevenLabs and the service was wound down. The code is still on GitHub, and a community-maintained self-hosted version exists, though it requires PostgreSQL, Elasticsearch, and several microservices to run.

What is Fabric?

Fabric is an AI workspace that combines file storage, note-taking, search, tasks, collaboration, and publishing. The Fabric Memory Engine automatically extracts, enriches, and maps relationships between everything you save. It handles the same content types Omnivore did (articles, PDFs, newsletters) and then keeps going: video, audio, images, slides, spreadsheets, design files. Where Omnivore was a place to read things later, Fabric is a place to keep, find, and work with everything.

Key differences

Omnivore was a reader. Fabric is a workspace.

Omnivore did one thing well: save web content and make it readable. That's a useful function, but it was the whole product. Your articles lived in Omnivore. Everything else lived somewhere else. Your notes were in Obsidian or Notion. Your files were in Google Drive or Dropbox. Your tasks were in another app. Omnivore was one more silo in a collection of silos.

Fabric replaces several of those silos. Articles, documents, meeting recordings, design references, links, videos, emails. They all live in one place, and the AI understands them all. You don't save an article to Fabric and then go somewhere else to do something with it. You save it, search it, reference it in an AI conversation, annotate it, share it, or publish it. All without leaving.

Search

Omnivore had full-text search and saved search queries. For a read-later app, it was good. But it was keyword search. If you couldn't remember the right words, you couldn't find the thing.

Fabric searches by meaning. Semantic search finds content even when you describe it differently from how it was written. Visual search lets you drop in an image and find similar content. Colour search finds assets by palette. You can search inside PDFs down to the page, inside slide decks down to the slide, and inside video and audio down to the timestamp. And cross-platform search pulls results from Google Drive, Notion, and Dropbox alongside your Fabric library. One search bar, everything you have.

AI

Omnivore had no AI features. You saved content, you read it, you highlighted it. That was the loop.

Fabric has a built-in AI assistant that knows your content. It can answer questions about your saved material, summarise documents, transcribe audio and video, generate meeting recaps, run web searches, and take actions inside the app. You can reference specific files or folders as context in a conversation. The AI gets smarter the more you save because it maps relationships between your content automatically. If you were using Omnivore alongside a separate AI tool like ChatGPT or Claude, Fabric collapses that into one place.

Text-to-speech

This was Omnivore's standout feature. The TTS voices were genuinely good, with multiple options including OpenAI voices. The voice would even change when reading block quotes. It was better than anything else in the read-later category.

Fabric doesn't have a dedicated article TTS reader in the same way. It handles audio and video transcription and AI-generated summaries of recordings and voice notes. If Omnivore's TTS was your primary reason for using the app, this is an honest gap. ElevenLabs launched ElevenReader using the same team, so that's worth a look if TTS is the thing you miss most.

Highlights and notes

Omnivore let you highlight text and add notes while reading. Those highlights could sync to Obsidian via a plugin, which was the core of many people's reading workflow.

Fabric has a full markdown editor with version history, embedded file references, and real-time collaborative editing. It's more than a highlighting tool. But if your specific workflow was "highlight article in Omnivore, sync to Obsidian, review in Obsidian," you'll need to rebuild that pipeline. Fabric integrates with Obsidian-adjacent tools through its API, MCP integration, and Zapier, but it's a different architecture.

Organisation

Omnivore used labels and filters. Simple, effective for a reading list.

Fabric uses Spaces (project containers), folders, tags, and multiple views including kanban, grid, list, and detail. It's designed for people managing more than a reading queue. If you're organising research projects, client work, creative references, or team knowledge, the structure scales. If you just want a clean reading list, it might feel like more than you need. That's fine. You can use as much or as little of it as you want.

The sustainability question

This is worth addressing directly. Omnivore and Pocket both shut down within eight months of each other. Two of the most-used read-later apps, gone. The read-later category has had a rough stretch, and it's fair to ask whether any tool you invest in will still be around in two years.

Fabric has a paid business model. It's not the most exciting thing to put on a comparison page, but it means there are customers, revenue, and a straightforward reason to keep building. That matters more now than it used to.

When to use each

Use Fabric if you want more than a read-later app. If you save articles, documents, meeting recordings, videos, and design references and want them all searchable, AI-queryable, and organised in one workspace, Fabric does that. It covers everything Omnivore did (minus dedicated TTS) and adds file storage, AI, collaboration, publishing, and search capabilities that Omnivore never had.

Use the Omnivore self-hosted fork if you're a developer comfortable running PostgreSQL and Elasticsearch, you specifically want an open-source read-later app, and you're okay maintaining it yourself. The community fork at omnivore.work exists. It's not actively developed by the original team.

Use Readwise Reader if your workflow is specifically about reading, highlighting, and syncing to a note-taking app like Obsidian. It's the closest feature-for-feature Omnivore replacement in the read-later category, at $9.99/month.

Use ElevenReader if Omnivore's text-to-speech was the killer feature for you. It's built by the same team at ElevenLabs. The feature set is narrower than Omnivore's was, but the TTS is best-in-class.

Why people move from Omnivore to Fabric

The patterns we hear from former Omnivore users:

They wanted to consolidate. Omnivore was one app in a stack of five or six. Articles in Omnivore, notes in Obsidian, files in Google Drive, tasks in Todoist, AI in ChatGPT. Fabric brings most of that into one place. Fewer apps, less context-switching.

They wanted AI on their content. Omnivore was great for saving and reading. Fabric lets you save, read, search, ask questions, summarise, and connect. The AI layer on top of your saved content turns a library into something you can actually work with.

They wanted durability. The Omnivore shutdown was a reminder that the tools you depend on need a reason to keep running. A lot of former users now weigh business model alongside features when choosing where to put their stuff.

They outgrew read-later. Saving articles is useful. But many Omnivore users were researchers, designers, students, and content creators who needed to do more with what they saved. Fabric's spatial canvas, collaboration tools, and publishing features support workflows that read-later apps weren't designed for.

FAQs

Can I import my Omnivore data into Fabric?

If you exported your Omnivore library before the shutdown (JSON or OPML format), you can upload those files to Fabric. If you didn't export before November 2024, that data is gone. There's no recovery option.

Does Fabric have text-to-speech like Omnivore did?

Not in the same dedicated article-reader format. Fabric handles audio and video transcription and AI summaries, but it doesn't have a built-in TTS reader that reads articles aloud. If that's your priority, ElevenReader (built by the Omnivore team at ElevenLabs) focuses specifically on this.

Is Fabric open source?

No. Fabric is a commercial product with a paid subscription. That means it's actively developed, maintained, and supported.

Does Fabric work with Obsidian?

Fabric has an API, MCP integration, CLI, and Zapier connection that can be used to build workflows with Obsidian and other tools. It's not a direct plugin like Omnivore had, but the integration surface is broad enough to support most workflows.

Can Fabric save articles from the web like Omnivore did?

Yes. The Chrome extension (web clipper) saves any page. You can also forward emails to your personal Fabric email address, sync a desktop folder, or save links directly from the app. Fabric automatically extracts and enriches the content of everything you save.

Is Fabric free?

Fabric does have a free tier with limited storage and AI.

Will Fabric be around long-term?

No one can predict the future of any product. What we can say is that Fabric has a paid business model with active customers, which gives it a structural reason to keep running and improving. We take the responsibility of holding people's content seriously.