Equerry vs Supermemory: personal assistant vs memory platform

If you landed here looking for a developer memory API, Supermemory is probably what you want. If you're a senior professional juggling client relationships — and you want an assistant that only ever sees the threads you hand it, shows you exactly what it kept, and briefs you before your next call — that's a different product. This page explains both.

If you're not building software, the alternative you want probably isn't another developer platform. Supermemory syncs from the cloud accounts you connect to it. Equerry is a personal assistant for iPhone: you forward or CC the client threads that matter, nothing reaches it that you didn't send, and it shows where an answer came from when it draws on your notes, emails or the web.

What Supermemory is

Supermemory describes itself as "the memory layer for AI agents" — a "context engineering platform powering enterprise APIs, developer plugins, and a personal app that remembers everything" (verified 2026-08-28 from supermemory.ai). So there are two things here: infrastructure other products build on, and a personal app anyone can sign up for, with a free plan and paid plans starting at $19/mo. You reach the personal side in a browser: save web pages, PDFs, images and audio; connect cloud accounts through connectors for Google Drive, Notion, OneDrive and S3; or wire it up over MCP (verified 2026-08-28 from Supermemory's documentation). On the hosted service that content is processed and stored on Supermemory's own infrastructure; the local install keeps it on your machine instead. The project's GitHub repository shows about 29.1K stars (verified 2026-08-28 from GitHub).

If you're a developer embedding a memory layer into your own AI product, or you want a browser-based tool that pulls from your existing cloud accounts, that's the shape Supermemory fits.

What Equerry is

Equerry is a personal assistant, built for the iPhone. You forward or CC it the client threads that matter, dictate a quick voice memo after a call, drop in a contract PDF or a screenshot, or save a typed note — and it remembers. Before your next client call it briefs you from what you've shared, and when you ask it something it answers with the source attached, so you can trust what you're about to repeat to a client. It's built for the person managing several client relationships at once, where one forgotten detail before a meeting is a real cost.

Two things set the trust level: what it ever sees, and what you can see of what it kept. You don't hand over a whole inbox or connect every cloud account — you decide, thread by thread, what Equerry gets. What you capture on the phone — the notes, the voice memos, the photos and PDFs — lives in your own iCloud account. A message you forward or CC takes a different path: it arrives at our email provider, which holds it briefly, and Equerry reads it there, turns it into searchable excerpts and extracted facts, and then deletes it from the provider. The raw message never lands in Equerry's database, and your own copy stays in your mailbox. Either way, every fact Equerry extracted is listed in Settings and deletable on tap, and deleting a note removes the source and everything derived from it.

Equerry vs Supermemory at a glance

The Supermemory details on this page — its own description of itself, its plan prices, which connectors sit on which tier, its license, its star count, and what its README says about the local install — were read on 2026-08-28 from supermemory.ai, its pricing page, its documentation, and its GitHub repository. Products change — confirm the current details before you decide.

Server-side platform vs. iPhone-native assistant

The sharpest distinction between these two products is how much of your material each one ends up holding. Supermemory is a cloud platform: when you save a link, sync a Drive folder, or upload a PDF, that material is processed on and stored in Supermemory's infrastructure. Connect a Drive folder and the folder is in scope, not just the file you had in mind. That's the right model for developers building memory-backed products, and it's fine for consumers who are comfortable with cloud-hosted tools.

Equerry splits it, and it's worth being exact about where the line falls. What you capture on the phone — the notes you type, the voice memos you dictate after a call, the photos and PDFs you save — lives in your own iCloud account. Email takes a different route, because it has to: a message you forward or CC has to reach a server to be read at all. It arrives at our email provider, which holds it briefly; Equerry reads it, turns it into searchable excerpts and extracted facts, and once that has succeeded deletes it from the provider. The raw message never lands in Equerry's database, and the original stays in your own mailbox, under your control. Attachments are read for their text and the file itself is then discarded entirely — no copy on our servers, none in your iCloud. If a message arrives from a sender you haven't approved, we hold only its sender, subject and arrival time while we ask you what to do; decline, or say nothing for seven days, and it's deleted from the provider unread.

What does stay on Equerry's servers is the derivative layer: the searchable excerpts and the individual facts, drawn from email and from everything else you share. Every one of them is listed in Settings, deletable on a tap, and deleting a note removes the source and every derivative with it. We'd rather set that out here than have you discover it later; the full account is in our privacy policy, and it's worth ten minutes before any client material goes into any tool, ours included.

For a consultant handling confidential client strategy, or an advisor with sensitive financial correspondence, the question that actually bites isn't only where a file sits — it's how much ever leaves your hands, and whether you can get it back out. Equerry knows about four client threads if you forwarded four. It doesn't hold the inbox they came from, it can't reach the ones you didn't send, and what it did keep from them is listed for you to read and delete item by item.

The input model is different too. Supermemory connects to your existing accounts and pulls from them. Equerry waits for you to share — you CC your Equerry address and every reply on that thread lands automatically, you snap a contract, you dictate a note after the call. Nothing is swept up that you didn't explicitly put there.

Is Supermemory open source, and does self-hosting solve your problem?

Yes, for the part that question is really about. Supermemory's main GitHub repository is MIT-licensed, and the project ships a local install — a single binary that runs on your own machine, with an option its README describes as fully offline, so that nothing leaves that machine. The saved data sits in one folder you can back up or move (verified 2026-08-28 from the repository). The hosted service you sign up for at supermemory.ai is a separate product, with a free plan and paid tiers above it; self-hosting and the cloud plans are two different ways to use the same project.

Underneath that search, though, is usually a plainer question: how do I keep this material off someone else's server? Self-hosting is one honest answer, and if you're comfortable running a server, backing it up, and keeping it patched, it's a good one. It is also a job. Somebody has to own that machine on the Tuesday it stops responding, and if you're a consultant between client calls, that somebody is you.

Equerry answers that question from the other end, and not by pretending to be the same kind of thing. It is not open source, there is nothing for you to self-host, and it does run servers — what sits on them is the layer that makes your material answerable: the excerpts and the facts drawn from what you share. A forwarded or CC'd message is read at our email provider and deleted from it once those excerpts exist; it never lands in Equerry's database, its attachments are read for their text and the files then discarded entirely, and your own copy stays in your mailbox. What you capture on the phone — notes, voice memos, photos, PDFs — sits in your own iCloud. Stating that plainly is rather the point; a comparison page that implied otherwise wouldn't be worth reading. What's on offer instead is a different kind of assurance: it only ever holds what you deliberately sent it, every fact it kept is listed in Settings and deletable on a tap, and it shows where an answer came from when it draws on your notes, emails or the web — so you can check what it kept rather than take a license's word for it. That is a weaker guarantee than a machine in your own house, and a stronger one than a tool that syncs your accounts and shows you nothing.

How each one actually works, from a client email to an answer

Take one real message. A client writes to say the board review has moved to the 14th, and that a new person — not your usual contact — now signs off on scope. Six weeks later you need that detail, and you need to be right about it. Here's the path it takes through each product.

Supermemory. You either save the message yourself, through the web app or an upload, or you connect the account it lives in and let it sync. The $19/mo plan carries connectors for Google Drive, Notion, and OneDrive, with Gmail and Granola on the tier above it and the rest — GitHub, S3, a web crawler — on the one above that (verified 2026-08-28 from Supermemory's pricing page). Either route, the content is processed and stored on Supermemory's own infrastructure — unless you're running the local install, in which case it stays on your machine. When you search, or when an app you've built queries the API, it retrieves the relevant pieces and returns them.

Equerry. You CC your Equerry address on your reply — once. Every later message on that thread lands too, so the client's follow-up two weeks on arrives without you doing anything. That message arrives at our email provider and Equerry reads it there — no other thread in your inbox is reachable, because none of them were sent. What it keeps is the specifics: the date that moved, the new approver's name, what was agreed, held as searchable excerpts and facts on our servers. The message itself is deleted from the provider once that has succeeded; it never lands in Equerry's database, and your own copy stays in your mailbox. Six weeks later you ask what the client said about the timeline, and the answer comes back with a source chip on it — the excerpt from that email — so you can read the sentence yourself before you repeat it on a call. And when the meeting is on your calendar, Equerry can hand you a brief beforehand, assembled from everything you've shared about that client — how meeting prep works covers that end of it.

The mechanical difference is what decides which one fits you. Under a connector model, what the product knows is a function of what your connected accounts contain. With Equerry, it's a function of what you chose to send it. Neither is better in the abstract; they're answers to different worries.

Supermemory vs mem0, Zep, and OpenMemory — and where Equerry sits

If you're weighing Supermemory against mem0, Zep, or OpenMemory, you're shopping one shelf: developer infrastructure. mem0 calls itself "drop-in memory infrastructure for AI agents and apps" and lists OpenMemory among its own projects; Zep leads with "agent memory, at enterprise scale" (both verified 2026-08-28 from their own sites). They're memory layers you wire into something you're building — an API, an SDK, an MCP server, a store you write to and query. Choosing between them is an engineering decision about recall quality, latency, self-host story, and cost at your volume, and it deserves a benchmark against your own data rather than a landing page. If that's your question, pick one of them. This page won't help you.

Equerry isn't on that shelf. It isn't infrastructure you build on — it's an assistant you use as it comes, on your iPhone, with nothing to wire up and no code to write. That makes it the wrong answer for the person shipping a product, and the right one for the person on the other side of the table: the one with four client relationships in flight and a call at nine tomorrow morning.

One disambiguation while you're in this category, because the names collide. mem0 describes itself as "drop-in memory infrastructure for AI agents and apps" (verified 2026-08-28 from mem0.ai); Mem is an AI notes workspace (verified 2026-07-10). Different products, different shelves. If it was the second one you had in mind, we cover it separately in Equerry vs Mem.

So the sorting question isn't which of these remembers best. It's which side of the table you're on. If you're building the product, take the infrastructure. If you're the one keeping track of several clients at once and getting ready for the next call, that's the job Equerry was built for.

When Supermemory makes sense

If you're building an AI product and need a battle-tested memory layer or API, Supermemory is designed precisely for that. And if you want a browser-based memory tool that syncs from your existing cloud accounts — and you're comfortable with server-side storage — its personal app is built for that shape.

When Equerry makes sense

If your work runs on email threads, client calls, and PDF contracts — and you need an assistant that holds the context of several relationships, never sees more than you handed it, shows you exactly what it kept, and hands you a briefing before each call — Equerry is built for that. The three deciding questions: do you want to choose thread by thread what your assistant ever sees, do you want memory you can verify and delete item by item, and do you want a briefing waiting before the call rather than a search box after it?

What the free tier covers, and what it doesn't

Supermemory has a free tier, and paid plans start at $19/mo, rising to $100/mo and $399/mo for higher usage and team features (verified 2026-08-28 from Supermemory's pricing page). Running the local install yourself carries no subscription at all. Equerry has no free tier: it's $39.99/mo, or $329/yr with a free first month for new subscribers on the annual plan.

We're not going to pretend that's a close call on price, because it isn't. If what you want is somewhere to save the articles and links you'll get to later, a free tier does that well, and you shouldn't pay twice as much for it. Buy the cheap thing.

The cost you're actually pricing is a different one. It's walking into a review having forgotten that the client's finance lead changed in March, or sending a follow-up that contradicts what you agreed on a call six weeks ago. For a consultant or an advisor, one of those in a quarter costs more than a year of any tool on this page. That's the comparison worth making — not $19 against $39.99, but the subscription against the mistake.

A four-question check before you put client material in either one

Rather than take anyone's word for it, ours included, these are the four questions worth answering about any tool before confidential client material goes into it. They take about ten minutes to check.

• Where does the raw file physically rest? Not the extracted summary — the actual email, the actual PDF. On your device, in your own cloud account, or on the vendor's servers?
• Can anyone but me read it? Who at the company can reach it, what happens to it while an answer is being composed, and is it used to train anything?
• Can I see, and delete, each individual thing it kept? Not a bulk wipe-the-account button — item by item.
• Does an answer show me its source? If it tells you the deadline is the 14th, can you open the email that says so?

Equerry's answers, plainly. The notes, voice memos, photos, and PDFs you save from the phone rest in your own iCloud account. A forwarded or CC'd email is read at our email provider and then deleted from it — the raw message never lands in Equerry's database, its attachments are read for their text and the files discarded entirely, and your own copy stays in your mailbox. What we hold on our servers is the derivative layer: the searchable excerpts and extracted facts drawn from everything you share — set out in full on our privacy page. What isn't there is everything else: no inbox connection, no account sync, so the only material in the system is material you chose to send. When you ask a question, the relevant material is sent to the providers that compose the answer — they're named on our AI disclosure page, and every one of them operates under terms that prohibit training on your data. Every fact Equerry extracted is listed in Settings and deletable on a tap, individually. And answers cite the specific note or email they came from, expandable inline, so you can check the source before you repeat it to a client.

For Supermemory's answers, read its current documentation and privacy terms rather than this page — and note that its two deployment routes give genuinely different answers to the first two questions. Products change, and a comparison page is no substitute for what a vendor says about itself today.

Last updated 2026-08-28.

All product names, logos, and brands are property of their respective owners. Equerry is not affiliated with, endorsed by, or sponsored by any other product mentioned; comparisons are provided for informational purposes.