How to actually find that restaurant someone recommended three weeks ago

Someone told you about a restaurant. Great food, you'd love it, it's in that neighborhood you're always in. You nodded. Maybe you said "send me the name." Maybe they did. Maybe you saved it somewhere.

Three weeks later, you're standing in that exact neighborhood, hungry, and you can't remember the name. You can't remember who told you. You vaguely remember it was Italian. Or maybe Japanese. It was definitely on the left side of something.

You open Google. You type "good Italian restaurant" and the name of the neighborhood. You get Tripadvisor lists, sponsored results, and places you've already been. The restaurant your friend recommended is buried somewhere on page two, if it's there at all. You pick something random. It's fine. Not great. You eat, pay, leave, and two days later you remember: it was the Japanese place. And it was on Instagram, not in a text.

This is a small problem. But it happens constantly. Not just with restaurants. With everything. Movie recommendations from friends go the same way, and so do TV show recommendations the second you hear them.

Why do recommendations disappear so fast?

Modern life runs on recommendations. A friend texts you a book. A colleague mentions a tool in a meeting. Someone shares a link in a group chat. You see a product on Instagram. Your sister sends a recipe on WhatsApp. A podcast host mentions another podcast. You overhear a movie title at a coffee shop.

Each one is a small gift: someone filtered the world for you and said "this is worth your time." And each one tends to slip by the same day, because recommendations arrive everywhere and live nowhere. Most casual facts fade fast unless you deliberately store them somewhere you can find again.

Think about where your last ten recommendations came from. A text thread. A Telegram group (and forwarding to yourself in Telegram doesn't work either). An Instagram story (gone in 24 hours). A Slack channel. A verbal conversation. An email buried under fifty others. A Reddit comment you scrolled past. A tweet in a timeline that refreshes every time you open it. That's typically eight to ten separate apps, none of which share a search box.

Now think about where you'd look for them. There's no single answer, because there's no single place. Every recommendation lives in the context where it arrived, and that context is impossible to search.

You can't search a conversation for "that thing someone mentioned." You can't search Instagram stories from three weeks ago. You can't search your memory for "I think it was Anna, or maybe it was Misha, and it was either a restaurant or a bar." This pattern repeats every time someone tells you about a great dermatologist or a friend recommends the perfect book 3 months before you go looking for it.

Why do existing tools fail at this?

You could save every recommendation manually. Copy the name, paste it into a note, add who recommended it and when. Some people do this. They have spreadsheets of book recommendations, Google Maps Saved Places lists of restaurants, Notion databases of movies. In our own attempts, and in conversations with friends who have tried it, these setups tend to fall apart within a few weeks: the capture cost beats the perceived value the second the conversation moves on.

The problem is time and friction. You're in the middle of a conversation when someone mentions a place. Are you going to stop, open an app, create an entry, type the name, add context, save, and then return to the conversation? Maybe once. Not twenty times a week. And the recommendations that come verbally (which are often the ones that matter most) never make it into any system at all.

Google Maps lets you save restaurants to lists, but you have to search for the restaurant first, which means you already need to know the name. And Google Maps doesn't save who recommended it, why, or what they said about it. It's a pin on a map with no story.

Apple Notes, Google Keep, or any simple note app could work as a running list. But finding "that Italian place" in a note titled "random stuff" six months later requires you to remember exactly what you wrote. You probably typed "Italian restaurant Anna" or maybe just "restaurant" or maybe just pasted a link with no context. Good luck searching for that.

Browser bookmarks are even worse. A bookmarked Google Maps link looks like every other bookmarked Google Maps link. No preview. No context. No memory of why it's there.

What does "remembering" actually look like?

What should happen instead.

dEssence is a free app: memory you don't have to maintain. Save it, forget it, ask for it later. No folders, no tags, no organizing. Capture with the Chrome extension, the Telegram bot, or the web app at dessence.ai, whichever is closest at the moment.

dEssence reads the message, extracts the restaurant name, notes who sent it and when. It pulls in details from the web: location, cuisine, hours, photos. You contributed one tap.

Three weeks later, you're in that neighborhood. You don't remember the restaurant. You don't need to.

You ask in your own words. You open dEssence and type "that restaurant someone recommended." Not "Italian restaurant Kreuzberg" (because you forgot it was Japanese and in Mitte). Just "that restaurant someone recommended." dEssence finds it because it reads what you saved, not just the words you typed. It shows you the name, the location, who recommended it, and the original message. Context restored without scrolling for it.

Or: you're already in the neighborhood and you remember you'd saved something for around here. You ask, "what did anyone recommend in Mitte." The list comes back. You pick one. You walk over.

That second move is what changes behavior. You don't need to remember the name. You need a memory you can ask in plain language, in the moment, without a search box that demands exact words.

Beyond restaurants

The same pattern applies to everything people recommend to you. It's also the same dynamic behind the promise you made in an article three weeks ago and forgot about.

Books. Someone mentions a title at a dinner party. You save it to dEssence with a voice note or a quick text. Two months later, you're in a bookstore and you ask, "what books did people recommend this fall." The list comes back with "Breath" by James Nestor and a note that Misha recommended it after your conversation about running.

Travel. Over six months, you save a dozen recommendations for a city you're planning to visit. Hotels from a colleague. Restaurants from Instagram. Neighborhoods from a blog post. Walking routes from a friend. When the trip approaches, you ask dEssence "everything about Lisbon" and get a single view of everything you've collected from every source. No spreadsheet. No Pinterest board. No "where did I save that?" That's also the move that turns a vague travel bucket list into an actual itinerary.

Products. You see a backpack on Instagram. You save the post. A week later, you're researching backpacks and ask dEssence "that backpack I liked." It shows you the exact post, the brand, the price, and similar items you've saved before. You make a decision based on your own curated references, not on Google's sponsored results.

Movies and shows. Friday evening. Four people. Everyone suggests something different. Instead of scrolling Netflix for forty-five minutes, you open dEssence and ask "movies people have recommended to me." You see a list with context: who recommended each one, why, when. You pick one in two minutes. You actually watch something good.

Recipes. You save a recipe from Instagram in March. In October, you're cooking for friends and want something impressive. You ask "that pasta recipe with the pistachio thing." dEssence finds it even though you didn't save the exact name, because it read the content and understood what was in the recipe.

How does this work under the hood?

dEssence does three things that traditional bookmarks and notes can't.

Content understanding. When you save a link, dEssence doesn't just store the URL. It reads the page, extracts key details (cuisine type, location, price range for a restaurant; genre and director for a movie; ingredients for a recipe). This means you can search by meaning, not by keyword.

Context preservation. dEssence remembers how things were saved. Who sent it. What they said about it. When you received it. Which conversation it came from. This context is what makes search feel natural. "That thing Anna sent" works because dEssence knows Anna sent it.

Retrieval in your own words. Instead of needing the exact title or filename to find something, you describe it the way you'd describe it to a friend: "that restaurant Anna sent me last spring," "the backpack from Instagram," "books recommended this fall." The difference between a filing cabinet and a friend who answers when you ask, "hey, what was that place you told me about?"

What does "not remembering" actually cost?

This might seem like a small convenience. A nice-to-have. But add it up.

Every week, you lose a few recommendations. A restaurant becomes "I think someone told me about a place but I can't remember where." A book becomes "what was that title?" A product becomes "I saw this thing somewhere but I can't find it." A recipe becomes "I saved it but I don't know where."

Over a year, those small losses add up to dozens, maybe hundreds. Each one is a tiny missed connection between your past self (who found something valuable) and your present self (who could use it right now). The loss is invisible because you never know what you missed. But the feeling is real: a vague sense that you've seen something good and can't get back to it.

Good recommendations are a form of filtered knowledge. Someone you trust spent their time and attention finding something valuable, and they shared it with you. Losing that is losing the benefit of your entire social network's taste and experience.

Frequently asked questions

How do I save restaurant recommendations from friends?

Forward the message (text, Telegram, WhatsApp, Instagram DM) to a memory app the second you receive it. Don't categorize, don't tag, don't open another app. The whole point is that the capture has to be as cheap as ignoring the recommendation, or you won't do it.

How do I find a restaurant someone recommended without remembering the name?

You need a search that understands meaning, not keywords. A memory app like dEssence lets you ask in your own words ("that Italian place Anna mentioned in spring") and finds it because it stored the sender, the cuisine, and the date alongside the link. Google Maps and Apple Notes were built for different jobs: map pins and quick notes, not contextual recall by description.

What's a workable way to remember restaurant recommendations?

Capture in the moment, retrieve by description. The first part is one-tap forwarding to a single place. The second part is natural-language search: you ask "that Italian place Anna mentioned in spring" or "what did anyone recommend in Mitte," and the saved item comes back with the original context.

Can Google Maps remind me about saved restaurants nearby?

Google Maps does have a nearby-notification feature for saved places. A more durable pattern is asking your own memory in plain language when you're hungry in a neighborhood: "what did anyone recommend in Mitte." The answer is on demand, not pushed.

Getting started

If you want to try this, the simplest version takes thirty seconds.

Worth saying plainly first: dEssence is in beta, the paid tier isn't finalized, and there's no native iOS or Android app yet. Capture goes through the Chrome extension, the Telegram bot, or the web app at dessence.ai. Search quality grows with what you've saved, so a near-empty account won't feel like much for the first few entries.

Install the Chrome extension, find the dEssence bot on Telegram, or open the web app at dessence.ai. Send it anything: a link, a name, a note, a forwarded recommendation. That's your first saved memory.

Next time someone recommends something, forward it. Don't think about categories or tags or where it goes. Just forward it and move on.

When you need it back, ask in your own words. "That restaurant." "Books people recommended." "The thing about Berlin." dEssence reads what you saved and answers the way you asked.

The shift is small but real: the recommendation stops being a thing you have to remember and becomes a thing you can ask for. That's when the tool stops being a tool and starts feeling like memory.