@polygres

our open source repos are available on GitHub! run by @wondersearchai

polygres.com
Joined September 2026
  • Tweets0
  • Following1
  • Followers6
Include

Only show posts containing:

Exclude

Hide posts containing:

Time range
-
Minimum likes
Just search
dale
@daleverettOct 15
For centuries, sailors didn’t change the sky. They built an instrument to read it. Your data doesn’t need changing either.
2
137
This is crazy! If you’re not using Wondersearch (SOTA Sparse Retrieval) you’re not living life
dale
@daleverettOct 14
AI search should just feel like finding the answer. WonderSearch (YC F26) lets your agents semantically search your docs directly: no embeddings, no vector DB, nothing to build or maintain. $0 when idle. Plug any agent in via MCP, API, CLI or SDK. Link below👇
4
177
The greatest way to search in my opinion
dale
@daleverettOct 13
Introducing WonderSearch 1.1 (YC F26): SOTA sparse retrieval, with zero neural ingestion. We beat Elastic’s ELSER on every BEIR benchmark they’ve published, by up to 13 points. No clusters. No embeddings. $0 when idle. Search without limit. Or the setup. 👇
2
3
183
Bad luck is a skill issue atp ngl
dale
@daleverettOct 11
Nobody is lucky. Some people just have better search. Introducing WonderSearch (YC F26) by @polygres backed by @ycombinator @fdotinc (link below)
1
3
207
We have so much more. This is just the start
dale
@daleverettOct 7

We built semantic search that doesn’t require pre-embedding your data (YC F26)

Article · @daleverett
1
5
347
we just killed Pinecone Weaviate Qdrant Milvus Chroma LanceDB Vespa FAISS Redis Elasticsearch OpenSearch my bad frfr
dale
@daleverettOct 4
We accidentally made embedding search obsolete. We just beat every dense embedding model we tested on SciFact by rethinking search completely. Wondersearch 1.1 is coming soon 🔎 More info in first comment 👇
3
5
283
Yo wait what
dale
@daleverettOct 4
We accidentally made embedding search obsolete. We just beat every dense embedding model we tested on SciFact by rethinking search completely. Wondersearch 1.1 is coming soon 🔎 More info in first comment 👇
1
3
262
We hit 120k views on our launch a few days ago… and a ton of feedback. We’ve been cooking and 1.1 is dropping very soon. Best part? Just switch the model 😉 No embeddings required.
dale
@daleverettOct 3
What happens when your data changes faster than your search can keep up? WonderSearch 1.1. Search your data as fast as it changes. No embeddings required +7.9% average benchmark gains Coming soon to @polygres (YC F26) link in first reply ↓
1
6
309
Better call dale
dale
@daleverettOct 1

Search Candidate Data Without the Embedding Hangover

Article · @daleverett
2
510
we just surpassed 1k followers!
2
8
436
New system one model for search. This is crazy
dale
@daleverettSep 27
Week 3 of YC. We just made semantic search serverless. Until now, semantic search meant embedding every document, every row and every chunk up front, and paying for it before anyone ran a single search. We flipped that. With @polygres you connect your data, choose what's searchable, and search it by meaning. There's no giant embedding job first. You pay for tokens only when you actually search. It runs on our brand new model that matches embedding-model recall in our (early, tbc) internal benchmarks, without neural ingestion. We believe semantic search should work like serverless compute. Acess rolling out to Polygres users (link below) or dm me.
1
8
521
The old way: spend 12k embedding everything so the intern can search “refund” The new way: intern searches “refund”
dale
@daleverettSep 27
Week 3 of YC. We just made semantic search serverless. Until now, semantic search meant embedding every document, every row and every chunk up front, and paying for it before anyone ran a single search. We flipped that. With @polygres you connect your data, choose what's searchable, and search it by meaning. There's no giant embedding job first. You pay for tokens only when you actually search. It runs on our brand new model that matches embedding-model recall in our (early, tbc) internal benchmarks, without neural ingestion. We believe semantic search should work like serverless compute. Acess rolling out to Polygres users (link below) or dm me.
1
1
4
470
Not many people know but we’re open source!
Lim Damien
@damienheJun 1

pgGraph: Graph Superpowers for Postgres (Full Open Source, Rust)

Article · @damienhe
6
488
Replying to @thePM_001 @annalea_l
Thank you!
2
16
Replying to @k_taro56v2 @daleverett @daltonmeon @damienhe @ycombinator
@k_taro56v2 @daleverett @daltonmeon @damienhe @ycombinator
2
2
35
Stealth is rather hard when you don’t know where you are
dale
@daleverettSep 21
How to be ninja (YC Edition) 🥷
3
582
Make search light again
dale
@daleverettSep 20
Lightsearch: a System 1 model for search. Most semantic search makes you pay before you search. Lightsearch flips it: send a query, the model figures out how to search, pay only at query time. we’re in private beta and looking for design partners with hard search problems. make search light. dm me
1
13
2,492
Wait what?
dale
@daleverettSep 19
Pool Noodle (YC F26)
4
410
Why the fuck are you building a search engine?
dale
@daleverettSep 19
I can break your AI agent with a can of Monster. Link in first comment.
1
2
352
This is crazy holy shit
dale
@daleverettSep 19
I can break your AI agent with a can of Monster. Link in first comment.
1
2
315
Load more