Keep vs Omnivore

Compare Keep and Omnivore side by side on features, pricing, and the workflows each one is designed for.

Keep

Save anything from the web and get it back as markdown for AI agents or a simple reading feed.

Free, paid from $10/mo

  • Markdown output built for AI agents and MCP clients
  • Auto-sync from RSS, YouTube, X bookmarks, GitHub stars, and newsletters
  • Semantic search across everything you've saved

Omnivore

Shut down

Open-source read-later app with strong newsletter and markdown workflows.

Free

  • Completely free and open source
  • Strong newsletter-to-library workflow with per-user email
  • Synced with Logseq, Obsidian, and Notion for PKM workflows

Feature comparison

Here's how Keep and Omnivore compare across the features people actually look for. They share 14 core capabilities; the differences show up in what each tool focuses on.

FeatureKeepOmnivore
Capture and save
Browser extension
Mobile apps
Save from email
Save tweets
Save YouTube videos
Save GitHub stars
Save PDFs
Save files (docs, spreadsheets)
Save audio files
Save via API
Auto-sync sources
RSS auto-sync
YouTube channel sync
X bookmarks sync
GitHub stars sync
Newsletter inbox sync
Library and reading
Reader view
Offline reading
Full-text search
Semantic / AI search
Highlights
Notes
Tags
Collections
Public sharing
Full-text RSS extraction
AI and agents
Markdown export for AI agents
Bulk markdown export
MCP server
CLI tool
Claude Code skill
AI summaries
Public API
Import and export
OPML import / export
Pocket import
Instapaper import
CSV / JSON export
Send to Kindle

Pricing

Omnivore has shut down, so pricing below is for reference only. Keep is free, paid from $10/mo.

Keep

  • Free

    50 saved items lifetime, browser extension, MCP server, markdown export.

    Free
  • Pro

    1,000 saves per cycle, all sources, bulk exports, higher API quota.

    $10/mo
  • Max

    5,000 saves per cycle, everything in Pro.

    $25/mo

Omnivore

  • Free

    All features free; open source.

    Free

Strengths and weaknesses

Both tools do their category well, but the specifics differ. Here's what each one is good at and where it tends to fall short.

What Keep does well

  • Markdown output built for AI agents and MCP clients
  • Auto-sync from RSS, YouTube, X bookmarks, GitHub stars, and newsletters
  • Semantic search across everything you've saved
  • Public API and Claude Code skill from day one

Where it falls short

  • No native mobile apps yet
  • No highlights or annotations
  • No Instapaper import yet
  • No Kindle send

What Omnivore did well

  • Completely free and open source
  • Strong newsletter-to-library workflow with per-user email
  • Synced with Logseq, Obsidian, and Notion for PKM workflows
  • GraphQL API returned markdown, friendly to integrations
  • Active community and regular updates prior to shutdown

Where it fell short

  • Shut down in 2024 after acquisition by ElevenLabs
  • No path to import back into a hosted version
  • Self-hosting requires non-trivial infrastructure

Which one should you pick?

Omnivore is no longer an option

Omnivore has shut down and is no longer available. Any of the active alternatives is a safer bet.

About Keep

Keep is a save-anywhere tool built around one idea: everything you capture should be available as clean markdown that an AI agent can read. Articles, tweets, YouTube videos with transcripts, GitHub stars, newsletters, RSS, and plain URLs all land in the same searchable library. People read their library in a clean in-app feed. Agents read it through the API, CLI, MCP server, and Claude Code skill, so Claude Code, Cursor, ChatGPT, and other tools can work against the library directly. Auto-sync pulls from RSS, YouTube, X bookmarks, GitHub stars, and newsletter inboxes on a schedule, so the library stays current without manual work. Semantic search runs across everything you've saved.

About Omnivore

Omnivore was a free, open-source read-later app that did everything right on paper: RSS feeds, newsletter inbox, PDFs, highlights, labels, filters, rules, full-text search, a GraphQL API that returned markdown, and sync with Logseq, Obsidian, and Notion. It ran on iOS, macOS, Android, web, and extensions for Chrome, Firefox, Safari, and Edge. It shut down on November 15, 2024 after ElevenLabs acquired the team for their ElevenReader TTS product. The cloud service deleted all user data; the open-source codebase still lives on GitHub for anyone who wants to self-host.

Keep exploring