Omphalis Review (2026): Pricing, Features & Honest Verdict

Omphalis turns articles, PDFs, podcasts and videos into structured reads rather than summaries. It maps how a source is built, explains hard passages in place, and links what you marked across your library. Best for: researchers and heavy readers who want to keep what they read. Price: from $9/mo (free plan: yes). Rating: 7.5/10.

What is Omphalis?

Most read-it-later tools now bolt an AI summary onto the top of a saved page. Omphalis refuses to do that, and says so repeatedly. Its position is that a summary is a way of not reading something, and that the useful work is making a hard source easier to enter: clutter stripped, layout untangled, two-column PDFs pulled into one reading order, transcripts cleaned, structure laid out, and the dense paragraphs explained where you actually meet them.

Everything goes through the same pipeline. Articles, PDFs, EPUBs, Word documents, podcasts, YouTube, newsletters and your own voice notes. You can subscribe to Substack or any feed and it lands in an inbox alongside your podcasts. It launched on Product Hunt in late July from solo maker Stanly Thomas, with a Chrome and Firefox extension, an MCP server so you can query your library from ChatGPT or Claude, and a founding-member offer of 90 days of full Premium with no card.

What Are Omphalis’s Key Features?

Structure, not the gist

The flagship idea. Instead of compressing a piece into three bullet points, Omphalis shows you how it is built: sections, argument flow, transitions, dense moments, claims, evidence, and the places worth re-entering. On a New Yorker feature it will mark the narrative sections, the quotes and the pivot points before you have read a word. This is the feature that justifies the product existing, and it is the one that is hardest to fake with a prompt.

Annotations that are first-class objects

When you mark a paragraph or a podcast timestamp as important, confusing, worth using or worth questioning, that mark is stored as an object with meaning attached, not as a highlight colour. Omphalis’s framing is that it preserves what you noticed rather than what a model guessed mattered, and the whole connection system downstream is built on those marks rather than on generic document similarity.

Cross-content connections

The Premium feature. Omphalis finds where one thing you saved supports, extends, updates or contradicts another, and builds those threads from the moments you marked rather than from every file you ever uploaded. It compounds: a piece you read in March can be picked up or challenged by one you read in August. Read the small print, though, because there is a caveat in it that we cover below.

Spaces and Memory

Spaces are separate shelves for separate parts of your life, so a work project and a personal interest do not end up in one pile, and nothing crosses between them unless you move it. Memory follows the topics you keep returning to, and the product explicitly says you can see every entry and forget any of it. That is a better privacy posture than “we use your data to improve your experience,” and it is one line rather than a policy page.

MCP and the ChatGPT integration

Omphalis exposes your library over the Model Context Protocol, so you can query it from ChatGPT, Claude or your editor without copying anything across. Free accounts get read-only access. Pro and above can read, save and add notes through the same channel. Scholar adds a library ingest API for your own scripts. Full-text search and data export are on every tier, including free.

How Much Does Omphalis Cost?

Omphalis starts free and the first paid tier is $9 per month, or $90 a year. Every annual plan saves 17%.

Free gives 100 comprehension units a month, which the site converts to roughly 40 articles or 6 hours of podcasts, plus 100 subscribed sources and 2GB of storage. Pro is $9/mo or $90/yr for 300 units (around 120 articles, 20 hours of podcasts, or a dozen PDFs), 1,000 sources and 25GB. Premium is $19/mo or $190/yr for 600 units (240 articles, 40 hours, 30 PDFs), unlimited sources, 50GB and cross-content connections. Scholar is $39/mo or $390/yr for 1,200 units (480 articles, 80 hours, 60 PDFs), 100GB, the ingest API and unlimited assistant answers. If you have a heavy month, unit packs start at $10 for 200 units.

Founding members get 90 days of full Premium with every feature and no card, then drop to Free unless they pick a plan. Once founding spots fill, new accounts get a 30-day Premium trial instead. Either way, nothing you saved goes away when the trial ends, and the library is kept forever on every tier.

Against the field, $9 for Pro is competitive. Readwise Reader sits in the same range, and most AI research assistants either cost more or meter you in tokens you cannot reason about. Which brings us to the thing Omphalis does better than almost anyone.

Does Omphalis Define Its Metering Unit?

Yes, and this deserves a section because so few products manage it. We have reviewed a run of tools this year whose pricing pages sell you credits, tasks or units and then never say what one is. Omphalis puts a definition inline, next to the number, on every tier: a comprehension unit meters “cleaning a source up, laying out its structure, and explaining the hard bits,” one unit is roughly one article, and a long PDF or a two-hour podcast costs more because it costs them more.

Then it converts each allowance into things you recognise. Not “600 units” but “around 240 articles, 40 hours of podcasts, or 30 PDFs.” It also states what is free: re-reading things you already saved never costs a unit, and voice notes are never metered. Everything draws from one monthly pool covering reading, transcription and listening.

The honest caveat is that “a long PDF costs more” has no published multiplier, so a Scholar user working through 60-page papers cannot calculate their real ceiling in advance. That is a genuine gap. But it is a small one next to an industry standard of publishing a number with no unit attached at all, and it is the single clearest signal on the site that somebody sensible is running this.

Who is Omphalis Best For?

Use Omphalis if you save far more than you read and want that to stop being a source of guilt. Use it if your inputs are hard: dense papers, two-column PDFs, two-hour interviews, technical talks full of terms you do not have yet. That is where the structural mapping earns its keep, and the site is candid that it is not worth much on a source you would have breezed through anyway. Use it if you want your reading queryable from ChatGPT or Claude without shipping it into a chat window by hand. And use the free tier as a real trial, because 100 units a month is about 40 articles, which is enough to know.

Skip Omphalis if you actually want summaries, because it is built to prevent that and you will find it slow. Skip it if you read mostly on a phone, since the product is browser-first with a Chrome and Firefox extension and no mobile app is advertised. And skip Premium specifically for cross-content connections until you have a large marked library, because the feature is explicitly described as working hardest on the hundred pieces you actually spent time with, not on everything you ever saved. That is an undisclosed practical cap on the headline reason to upgrade, and it is buried in the plan detail rather than stated up front.

Best Omphalis Alternatives

Readwise Reader is the mature competitor and sits in a similar monthly range. Far better on mobile, better on highlight export and spaced repetition, and much weaker on structural analysis of a difficult source. If you mostly want a good read-it-later app with highlights that go somewhere, take this.

NotebookLM is free and very good at question-answering across a document set, with audio overviews that many people love. It is a research assistant rather than a reading environment: no clean-read pipeline, no annotation objects, no library that persists as your own thing.

Obsidian with a web clipper costs nothing, keeps everything in local Markdown, and gives you total control. It also gives you no comprehension layer whatsoever, so you are trading Omphalis’s whole value proposition for ownership and flexibility.

Final Verdict: Is Omphalis Worth It?

Omphalis is the rare tool with a real argument behind it. The refusal to summarise is a position, not a limitation, and the product is built consistently around it: structure before gist, marks as first-class objects, connections that grow from what you chose rather than from cosine similarity between files. The pricing page is the best we have read this year, defining its unit inline, converting every allowance into articles and hours, and stating plainly what is never metered. The library is kept forever on every tier and data export is on the free plan. Those are the decisions of somebody building a thing they intend to use themselves.

What holds it back is early-stage reality. One maker, no named customers, no testimonials, and a headline Premium feature that quietly works best on a hundred pieces rather than a whole library. The long-PDF multiplier is unpublished, so heavy users cannot forecast their ceiling. None of that is disqualifying. Start on the free tier, or grab a founding-member slot for 90 days of Premium with no card while they last, and give it the sources you have been avoiding. If the structural map changes whether you finish them, $9 a month is not a difficult decision.