Search that goesnarrow and deep.
Search the web without signing in, and read AI answers that cite their sources.
We aim to give deeper results when you come back to the same topic.
An answer is where you start, not where you stop.
Every answer on hiniki comes from someone's page, so we always show where it came from. We don't want search that keeps people to itself. We want search that sends people to the pages worth reading.
Today, every answer links to its sources. Next, we're shaping hiniki so more people open the original page, and the people who make the web get the visit.
Find it, read it, keep going.
Easy for people to read. Easy for agents to build on.
-
Web search
We crawl and index public web pages ourselves, and show you the results.
-
AI chat with sources
Answers built from the pages we found, with each source shown next to the answer.
-
Agent API
Compatible
/v1/searchand/v1/extract, with an issued key. Email us for access. -
HinikiBot crawler
Reads public web pages and follows each site's robots.txt rules.
Privacy and
records
You don't sign in, but that doesn't mean nothing is recorded. We keep search and request logs, including calls to search providers. Once a day, we remove user and session identifiers from records at least 9 months (270 days) old, and keep the query text, timestamps, and measurements to improve search quality. Your recent searches stay in your browser. AI conversations are tied to an anonymous cookie and stored on our server with an expiry set to 90 days after your last activity and up to 100 turns per conversation. Expired conversations are deleted in daily batches, so deletion may take longer when there is a backlog. To answer, we may send queries, questions, conversation context, and reference text to external search providers, and to AI providers through an AI relay. Those providers keep and handle data under their own policies.
Get in touch.
hiniki@hiniki.net- Business name
- hiniki
- Representative
- 강현진 (sole proprietor)
- Address
- 경기도 안양시 동안구 시민대로327번길 11-41, 안양창업지원센터 3층
Korean business registration no. 728-05-03736
Email us about the service, the API, or the crawler.
Site owners can stop HinikiBot by adding the rule below to robots.txt. It stops future crawling. For pages we've already collected, email us.
User-agent: HinikiBot
Disallow: /