I'm Building a Decentralized Amazon
Hi HN! Agora is an AI search engine for e-commerce products. We've indexed 21 million products sold on 62,000 different Shopify and WooCommerce stores and made them easy for customers to search and purchase. Think of a decentralized Amazon. I originally built Agora as a side project and launched it here on HN. The search took 5 - 10 seconds to return results and the app looked like it was designed in MS Paint (by yours truly). https://news.ycombinator.com/item?id=38635695 Since the MVP launch, here are a few technical challenges that we've been working on: 1. Figured out how…
What it does
In the maker’s words, at launch
Hi HN! Agora is an AI search engine for e-commerce products. We've indexed 21 million products sold on 62,000 different Shopify and WooCommerce stores and made them easy for customers to search and purchase. Think of a decentralized Amazon. I originally built Agora as a side project and launched it here on HN. The search took 5 - 10 seconds to return results and the app looked like it was designed in MS Paint (by yours truly). https://news.ycombinator.com/item?id=38635695 Since the MVP launch, here are a few technical challenges that we've been working on: 1. Figured out how to return results in under 300ms. We are using a self-hosted version of Meilisearch on a high memory Digital Ocean server. The app is built with Remix, specifically to optimize for pre-fetching of data. To power the AI search, we convert our products into both text and image embeddings. Text embeddings with Cohere and image embeddings with Clip. We are now working on further server optimizations and data sharding to get the search speed under 200ms. 2. Expanded the crawler to index stores built with Shopify, WooCommerce, Squarespace, and Wix. Additionally, we adapted the crawler to index major retailers including Walmart, Target, Macy's, and Nordstroms. We also built a "recrawler" that updates all stores on Agora every 24 hours. Then when a user clicks on a product, we do a live check of the variants and stock. This ensures that product data is always up to date. 3. Added deeper e-commerce query detection. We noticed that some searches are objective keywords like "camping tent" and others are subjective queries like "what is the best camping tent under $500". When a user searches on Agora, we pass the query through a text classifier to determine if it's subjective or objective, then change the AI tooling and UI accordingly. We served 2.1 million search queries in 2024 and are now using this data to get even more precise about query detection. My vision for Agora is to aggregate the world's e-commerce products into one place. I believe this will create market efficiencies for customers, developers, and merchants. I'd love your feedback!
Does the same job
all alternatives →
- III indexed 25,000 Shopify stores to build an embedded shopping API2025 · developer.searchagora.com · ▲9
Hi HN! I’m working on Agora, an AI search engine for e-commerce products. Excited to launch the Agora API that helps other developers embed shopping experiences into their app. We provide endpoints to search, purchase, and track products sold on 25,000 different Shopify and WooCommerce stores. Early customers are building social shopping apps, marketplaces, and AI shopping assistants. I’d love your feedback!
- IBI built a service to help companies save on their AWS bills2020 · ▲73
Hey HN: I'm Kaveh, the founder of Usage (https://www.usage.ai/) We help companies drive down AWS costs. Why? Because the way it's done now is a pain. Stakeholders, especially engineers, are required to spend unnecessary time manually finding underutilized or overly expensive EC2s. We believe the optimization process should be done automatically through a series of sophisticated algorithms. At the moment, there are over 70,000 AWS EC2 prices - doing that manually just won't scale at most organizations. My background is in software engineering. Previous to founding Usage, I…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, May 2025
the whole month →
- C9
Life & fun · 2025 · felixrieseberg.github.io
