Opyt vs Readwise

Readwise queries the articles you saved. Opyt queries everything the people you saved have ever published.

Readwise Reader is the best reading tool on this list, and it ships a hosted MCP server at mcp2.readwise.io that searches highlights and Reader documents with full text plus semantic matching, and writes as well as reads. The difference is not the pipe. It is what ends up in the store.

Reader's corpus is the articles you personally routed through Reader, a manual pipe with you as the pump. Opyt reads the saving you already did on X and Substack from your logged-in browser session, then expands each person you confirm into their whole archive: posts, repos and essays you never bookmarked.

At a glance

	Opyt	Readwise
What goes in	Whole archives of people you endorsed, plus standing queries	Articles you send to Reader
Who does the filing	Nobody. Your existing follows and bookmarks are the input	You, one save at a time
Sources	X, Substack, GitHub, blogs, arXiv, OpenAlex, any URL	Articles, PDFs, ePubs, newsletters, YouTube
Where it lives	One SQLite file on your machine	Readwise's cloud
Price	Free, MIT. Every source free; a few OpenRouter credits at ingest	Subscription
Reading experience	None; your AI client is the interface	Inbox, highlights, annotation, mobile, offline, TTS

Where Readwise is better

Reading itself. An inbox, highlighting, annotation, ePub and PDF, text-to-speech and a real mobile app. Opyt has none of it.
Spaced repetition. The daily review over your own highlights is a habit product. Opyt has no equivalent and no plans for one.
Stability. Readwise ingests through documented APIs. Opyt's X and Substack ingesters are browser-session scrapes, and X's pagination alone needs a 'no new item' terminator plus a seen-id set plus a 100-page backstop, because the cursor never ends.
Reach. Reader works on any device with a browser. Opyt needs an MCP client on a Mac or Linux machine.
Its MCP writes: create documents and highlights, add and remove tags, move between inbox, archive and shortlist, add notes, bulk edit metadata. Opyt's MCP never writes to a source.

Where Opyt is better

Documents per human action. One add_oracle confirmation ingests 3,412 posts, 38 repos and 96 essays in one pass, free. One Reader save ingests one article.
Sources Reader does not touch. The live store is 2,417 X posts, 267 GitHub repos, 28 blog archives, 24 papers and 2 Substack posts; Reader ingests none of the first two categories.
Counting over the whole store rather than a sample. One summer: 96 saves on tariffs and inflation, 84 on the grid and data centers, 41 on GLP-1 trials, and 1,551 of 2,804 from people you had confirmed.
One SQLite file. A full export of the live store is 115.9 MB and builds in 4.2 s, and read access is a link you send and revoke in one call.

Which one

Pick Readwise if

You want to read better, not just retrieve better
You read on a phone or a tablet
Highlights and review are the point

Pick Opyt if

Your bottleneck is what gets in, not what you can find
You want the rest of an author's work, not just the piece you saved
You live in Claude Code, Cursor or a similar client

Running both is the normal answer here. Readwise owns the act of reading, and what Opyt owns is getting the content in while you do nothing.

install Opyt free, MIT, five minutes

Opyt vs ObsidianNote apps and second brains

Opyt vs RecallAI second brains

all comparisonsevery category, side by side

useopyt.comwhat Opyt is for

hello@useopyt.comquestions and feedback