EarlGrid/brand-assets
Viewer • Updated • 1
EarlGrid Inc.
company
Verified
AI & ML interests
Speech AI, Named Entity Recognition, Text-to-Speech, Pronunciation Modeling, Cross-lingual Knowledge Graphs, Open Speech Data
Recent Activity
Organization Card
We put your name into the data AI learns from.
Speech AI learns from patterns in its training data. A name with non-English origin or a unique spelling is rare in those patterns, so voice assistants mishear it, mispronounce it, and send the customer who asked for you somewhere else. EarlGrid fixes this at the source, so that AI hears your name and says your name.
We construct the data infrastructure for names, so that AI can hear a name right, find it instantly, and understand what it actually points to. Engineers call these proper names 'named entities.'
What you get
1 Training data fix. Your name, in text and audio, published upstream in the open datasets speech AI learns from.* (Speech-to-text.)
2 Pronunciation data submission. A pronunciation record for every name, submitted to each voice assistant and lexicon that accepts one.** (Text-to-speech.)
3 An Alexa skill submission. An Alexa skill, with your name as its invocation name.
4 3 recognition reports. Same recordings and same method before the fix, at week 3, and at month 6.
* Mozilla Common Voice, Wikimedia Commons via Lingua Libre, and Wikidata. (+Wikipedia if the brand already qualifies)
** Wikidata (IPA), CMUdict and open lexicons, Alexa (via skill), Samsung Bixby, and Naver CLOVA. (+Siri if the brand has an iOS app)
How it runs
3 weeks of our work. Then the platforms take theirs.
1 Tell us the name. How you say it, and how you'll accept it being said. We record the baseline before we touch anything.
2 We build the records. Text, audio, and pronunciation in every format the platforms take: IPA, respellings, phoneme tags, and the data files AI learns from.
3 We submit them. To the open datasets and voice assistants on our list, with receipts.
4 We test it 3 times. Same recordings, same method: before the fix, at week 3, and at month 6. Results in writing.
Do names matter?
Proper names (named entities) often carry more meaning than any other word or phrase in a sentence. Take "EarlGrid was born in California." The words "EarlGrid" and "California" give you a more vibrant picture of the information than just "born" or "company" do. Or take "That person appeared in the movie Interstellar." The title Interstellar points to exactly one thing in the world, and that sharpness gives it more significant information than the other phrases and words such as "that person," "appeared," "in," or "the movie."
Why is AI bad with names?
Names come in many shapes from brand names and product names to titles of songs, films, series, and games as well as names of people, places, institutions, and companies. These are the words that carry the core information in human communication. And these are exactly the words AI gets wrong more frequently than we want. The rarer and the more distinct the name, the worse AI does with it. Here lies the paradox: a name exists to pick out one single thing in the world, so by nature it has to be unique. That very uniqueness pushes it outside the patterns in AI's training data, and AI struggles to recognize it.
Who is EarlGrid?
EarlGrid Inc. is a mission-driven company committed to building a cross-lingual knowledge graph of proper names. We are a global team, born in Silicon Valley with an Asia hub in Busan, Korea in June 2026.
What our founder saw
Machines have been struggling with names for as long as they have been listening. One founder spent nearly 2 decades noticing this up close. Early on, she transliterated geotags into Korean for a Fortune 500 company in the Bay Area, and watched place names break navigation for Korean users. Later, working on titles for a global streaming platform, she saw key names and phrases (a.k.a. KNPs) come through inconsistent across languages, and saw how much time and money it took to fix them one at a time. At some point, noticing was not enough.
The decision
So she closed the company she had given 9+ years of her life to, and started the one she needed. EarlGrid is that company, built slowly with people who care about getting things right. We build the cross-lingual named-entity knowledge base that speech AI should have had from the start, at the point where linguistics, speech AI, and content intelligence meet.
Our path
We fell often, and we still do. We just get up and keep walking, and the path is whatever is behind us.
What does the name 'EarlGrid' mean?
'Earl' (ì–¼) is the Korean concept of spirit, essence, and core identity. 'Grid' is the structural matrix that brings order.
We spelled the Korean word ì–¼ (romanized eol) as 'Earl' so English speakers could say it. That's the compromise every founder with a name from somewhere else makes, and it's the problem we work on. Ask your voice assistant for EarlGrid and see what it hears. The name is also the method. For every proper name we keep the 'eol' (earl ì–¼: essence and identity), what it means and how it sounds, and the grid, the structure that enables machines to find it and get it right.
Make your name easy to find in voice search
Speech AI Named entity recognition Text-to-speech Pronunciation modeling Knowledge graphs Cross-lingual NLP Open data
Silicon Valley HQ: San Jose, CA, U.S.A. · Asia hub: Busan, Republic of Korea
Home · Product · About · Contact · LinkedIn · YouTube · Instagram
models 0
None public yet