recall vs innernet
recall remembers what you consumed. innernet remembers what you decided.
the short version
recall remembers what you consumed. innernet remembers what you decided.
recall is a very good AI knowledge base for the things you read, watch and listen to — save from anywhere, get a summary, get tagged automatically, chat with it across several models, and get quizzed so it sticks. Hundreds of thousands of people use it and the execution is strong.
But a library of what you took in from the world is a different object from a record of your own judgment. The first tells an AI what you've been exposed to. Only the second tells it how you think.
what recall is building
"Save, summarize, and chat with your articles, videos, podcasts, PDFs, and notes — with ChatGPT, Claude, or Gemini." One-click saving from 15+ sources including YouTube and Spotify, automatic AI summaries with audio playback, smart tagging that organises without you, a multi-model chat interface, spaced-repetition quizzes for retention, and a connection map that links related ideas across everything you've saved. Browser extensions across every major browser plus iOS and Android. Free tier, premium above it.
The knowledge-graph-of-what-you-consume idea is well realised, and the retention layer on top is a smart addition most competitors skip.
what innernet is building
A personal memory layer that the AI tools you already use read from. npx innernet connects Claude Code, Cursor, Codex, Windsurf, Gemini CLI, Claude Desktop, VS Code, Cline, Zed and Continue to a single memory, with an MCP endpoint for everything else. netti — one per person — files what accumulates into Context Maps: dimensions that read as living documents, nodes for the people and decisions that recur, commits and branches you can work inside.
The material is your own reasoning, not the world's.
where they diverge
1. inbound vs. outbound memory
recall's corpus arrives from outside: articles, videos, podcasts. innernet's comes from inside: what you concluded, why, and what changed it. Both are memory. Only one of them can answer "why did we go with this architecture" — and that question comes up far more often than "what did that podcast say."
2. connections between sources vs. conclusions about you
recall's connection map links ideas across the things you saved — genuinely clever, and useful for research. But a link between two articles is a fact about the articles. innernet's dimensions are statements about your project and your thinking, with sources attached and a history showing when each one changed. The graph is about a different subject.
3. remembering it yourself vs. your tools remembering
The spaced-repetition layer says something about recall's goal: it wants the knowledge in your head. innernet has no interest in your recall — it wants your tools to arrive already informed so you don't have to hold it at all. Opposite theories of where memory should live, and both are defensible.
side by side
| recall | innernet | |
|---|---|---|
| Material | content you consume | reasoning you produce |
| Structure | auto-tags and a connection map across sources | dimensions and nodes about your work |
| Where memory should live | partly in your head, via spaced repetition | in your tools, so it doesn't have to |
| AI's role | chat over what you saved | writes and updates conclusions |
| Reachable by Claude Code / Cursor | not the surface it's built for | yes — one command |
| Change over time | the library grows | positions versioned, branchable, roll-back-able |
| Best question it answers | "what did that source say?" | "why did I decide that, and when did I change my mind?" |
where recall is genuinely better
- Capture breadth. YouTube, Spotify, podcasts, PDFs, articles, from any browser. We're much narrower.
- Summarisation and audio playback. Good at turning long material into something usable quickly.
- The connection map. Surfacing links between things you saved months apart is a real feature and a pleasant one.
- Spaced repetition. Almost nobody else bothers, and for actual learning it works.
- Scale and polish. A large user base and a mature product across every platform.
where innernet is stronger
- It holds decisions and reasoning, which is the memory that changes what an AI does next.
- It's read inside the tools you work in, without you opening anything.
- Change is first-class. Positions are versioned and branchable; the view you abandoned stays readable.
- Per-fact disclosure. Every fact carries how far it may travel, enforced when a tool reads.
- netti forms a view. Not a tidy library of what you saved, but a statement of what it adds up to.
which one you want
Use recall if you consume a lot and want it saved, summarised, connected and retained.
Use innernet if you want the AI tools you work in to understand your projects and your decisions without being told again each session.
Both is sensible. One is your reading; the other is your reasoning.
the question we actually get
"Recall already builds a knowledge graph of everything I've saved. Isn't yours just the same idea?"
Same technique, different subject. Their graph connects sources to each other. Ours connects facts to you — what you decided, on what evidence, and which earlier position it replaced. You could read every node in a content graph and still not know how the person using it thinks, which is exactly the thing an AI needs in order to be useful to them.
Accurate as of 3 September 2026, based on recall's public site. If we've described their product incorrectly, tell us and we'll correct it.
this is what innernet remembers. you can keep your own.
innernet is memory your ai tools read live — projects, decisions, the things you’d otherwise re-explain every session. we write these letters out of ours.