Rtzr
Seoul, KR · 1 known investors
ReturnZero (리턴제로) is a South Korean company developing AI-based speech recognition technology, positioned as an interface to help people interact with machines more naturally.
Also known as (주)리턴제로 · Return Zero Inc. · ReturnZero · RTZR · Vito AI · 리턴제로
Investors · 1
Company profile
researched Aug 2026Return Zero Inc. (Korean: 리턴제로, also branded RTZR) is a Seoul-based company building AI speech and language technology, centred on speech-to-text (STT) and text-to-speech (TTS) together with what it describes as full-stack AI engineering capability. Its portfolio spans three products: RTZR STT, a business-facing speech recognition solution available via cloud API and on-premises deployment; Callabo, a meeting-recording and analysis tool that captures and summarises online meetings held on platforms such as Google Meet and Zoom; and VITO, a consumer Android application that records phone calls and converts them into searchable text conversations.
RTZR STT is offered in batch/file and real-time streaming modes and includes speaker separation (up to ten speakers), custom vocabulary recognition, additional acoustic and language-model training, punctuation insertion and profanity filtering. The company markets Korean and Japanese recognition for enterprise use and publishes its own benchmark comparing RTZR STT against NAVER, Google and OpenAI on a one-hour Korean conversational speech dataset. A developer platform (developers.rtzr.ai) provides sign-up, a console for issuing secret credentials, token-based authentication, SDKs and tutorials in Go, Java, Kotlin, Python and Swift published under the vito-ai GitHub organisation, and a default free quota of 600 minutes; batch transcription results are retained for three days before deletion.
The company is registered as (주)리턴제로 / Return Zero Inc., led by CEO Lee Chamsol, with offices on the 8th floor of the Sindeok Building, 343 Gangnam-daero, Seocho-gu, Seoul.
Business model
Return Zero operates a mixed B2B and B2C model. Enterprise revenue comes from the RTZR STT speech recognition engine, sold both as a cloud API priced per second of audio and as an on-premises deployment, plus Callabo, a meeting transcription and analysis product sold for individual and corporate use. On the consumer side it distributes VITO, a free-to-install Android call-transcription app that also serves as a source of speech training data for the STT engine.
Usage-based pricing for the RTZR STT API charged per second of audio (marketed as 1.5x to 3x cheaper than competitors), enterprise/on-premises licensing, and subscription-style access to the Callabo meeting product for individuals and companies.
Traction
The company reports more than 1 million users of its VITO Android call-transcription app and about 15 million hours of accumulated speech data. Published deployments include Shinhan Bank's next-generation AI call centre, the Gwangju Fire Department, AICC vendor MindWareWorks for Japanese-language contact centre work, Channel Talk's sales team, and the Chun-Myung fortune-telling app, where the company reports a tripling of purchase conversion after adding transcription.
▸Full profile — market position, technology, go-to-market, geography
Market position
The company positions RTZR STT as the most widely used speech recognition engine in Korea and publishes internal benchmarks placing its Korean conversational accuracy and processing speed ahead of NAVER, Google and OpenAI, along with claims of 35% lower error rates than competitors generally, 48% lower error rates on noisy phone audio, 2.5x faster conversion and per-second pricing 1.5x to 3x cheaper. These figures are company-reported.
Stated differentiators are a large proprietary Korean speech corpus (~15 million hours) sourced from its own consumer app rather than licensed datasets, accuracy and latency on Korean and Japanese conversational and telephony audio, support for both cloud and on-premises deployment, security and stability credentials from adoption by a leading Korean financial group, and a developer-friendly API.
Technology
The core technology is Korean- and Japanese-language automatic speech recognition, including an end-to-end model, offered in batch/file and real-time streaming modes with sub-300ms streaming response, speaker separation for up to ten speakers, custom vocabulary recognition, additional acoustic and language-model training, sentence punctuation and profanity filtering; the company also states TTS capability. Models can be deployed in cloud or on-premises environments. Access is via an OpenAPI with token-based authentication, a developer console and gRPC-based client libraries and SDKs in Go, Java, Kotlin, Python and Swift. The company cites roughly 15 million hours of speech data gathered from its own VITO app's user base as its training data advantage.
Go-to-market
Enterprise sales through a contact-sales channel with published customer case studies, combined with self-serve developer adoption via the RTZR STT OpenAPI, a developer sign-up and console, documentation, published pricing and a 600-minute free quota, plus open-source SDKs and sample code on GitHub. VITO acts as a consumer distribution channel on Android.
Enterprises and public-sector organisations needing speech recognition — notably financial institutions and AI contact centres (AICC), customer-service and sales teams, mobile app and video-processing developers, and government agencies including fire departments — alongside individual consumers using the VITO call transcription app and business teams using Callabo for meetings.
Geography
Headquartered in Seocho-gu, Seoul, South Korea, with product focus on the Korean market; speech recognition is offered for Japanese as well as Korean, including Japanese-language AI contact centre deployments.
Compiled by commissioned research from 7 cited public sources — announcements, filings, and press listed under research sources below.
Key figures
latest reportedCompany-reported or press-reported figures, each dated to when it was claimed — not independently audited.
▸Research sources · 7
primary sources listed
- Rtzrrtzr.ai · web
7 public sources were cited for this profile; the first-party ones are listed here.
Frequently asked questions
- What does Rtzr do?
- Return Zero (ReturnZero) is a Seoul-based AI voice technology company offering Korean and Japanese speech-to-text and TTS.
- Who are Rtzr's investors?
- Rtzr's investors include Atinum Investment Co. Ltd..
- Where is Rtzr headquartered?
- Rtzr is headquartered in Seoul, KR.