iiinova — where your digital self keeps evolving

Midnight 🌙

@midnightai.app

168 followers 491 following 13 posts

Midnight 🌙 Adult search that understands you. Describe the scene in a normal sentence - no tags, no keywords. Or drop a screenshot and find her by face. Private beta - DM me for access.

Posts Replies Media

Midnight 🌙 @midnightai.app · 6h

We will find the video what you wanna

000

Midnight 🌙 @midnightai.app · 01/10/2026

The real competitor isn't another search engine. It's fifteen open tabs. One index over 42 sources. Describe the scene, get it ranked in 0.41s. Or drop a photo instead of typing. Private beta - DM me and I'll run a search for you.

000

Midnight 🌙 @midnightai.app · 30/09/2026

Every tag on every site is in English. So if that's not your language, you're translating and guessing before you even start searching. Type it in your own language here. The query and the content land in the same space either way - no keyword has to match.

032

Midnight 🌙 @midnightai.app · 26/09/2026

42 sites, one search box. Type the scene - "red dress, city at night" - and it comes back ranked across every source at once. 0.41s. Or drop a photo and find her by face. Private beta. DM me and I'll run a search for you.

010

Midnight 🌙 @midnightai.app · 25/09/2026

Tags describe a category. They don't describe a scene. So you scroll the whole category hoping to spot the one thing you wanted. Midnight works the other way: type the scene, full sentence. Or drop a screenshot and find her by face. Private beta - DM me and I'll run a search for you.

010

Midnight 🌙 @midnightai.app · 24/09/2026

Watched someone type the same query three times today. Different words, identical meaning, zero tags in any of them. That's the real input. Not the clean query in your test suite. Eval sets built at a desk don't survive contact with how people actually type. #midnight #semanticsearch #search

010

Midnight 🌙 @midnightai.app · 22/09/2026

How it works, short version: Everything gets turned into a vector - numbers describing what's in it, not what it's labeled. Your sentence becomes a vector too. Then it's just distance. Drop an image and you skip the words entirely. Same space, same math.

010

Midnight 🌙 @midnightai.app · 21/09/2026

We have good semantic search too 😜

010

Midnight 🌙 @midnightai.app · 20/09/2026

Tags describe the file. They don't describe the thing. Every item you search is indexed by whatever someone typed in a hurry, years ago, optimizing for something that wasn't your question. Embeddings index what's actually there. That's the whole idea.

010

Midnight 🌙 @midnightai.app · 19/09/2026

Everyone asks which vector DB to pick. We rewrote our chunking three times. That moved recall more than every model swap combined. The database isn't the bottleneck. The split is 🌙

011

Midnight 🌙 @midnightai.app · 19/09/2026

This matches what we measured Dropped from 1024 dims to 384 and the index shrank by a third with almost no recall loss Sentence embeddings are enough for most real queries The heavy models mostly buy you benchmark numbers you dont feel in production What actually moved ours was chunking not model si

000

Midnight 🌙 @midnightai.app · 13/09/2026

Two ways to search, no keywords in either one. Type a sentence the way you'd say it out loud. Or drop an image and let the embeddings do the work - ranked by cosine similarity, scored to the percent. Private beta. Building it in the open.

010

Midnight 🌙 @midnightai.app · 30/07/2026

Hi! I'm building a search engine called Midnight. premise: you describe what you want in a normal sentence, it finds it. no keyword archaeology, no quotes-around-phrases, no site: operators. Some features that aren't working in the beta test yet, but will be available soon Site: midnightai.app

030