What BrainSoft does

BrainSoft (Brainsoft Inc., 주식회사 브레인소프트) is a South Korean AI startup founded in August 2018 and headquartered in Gangnam-gu, Seoul, that develops high-resolution spectrogram-based voice AI technology. Its core proprietary "DJ Transform" algorithm — a patented frequency extraction method inspired by the human auditory system — achieves 1Hz frequency resolution and 1msec time resolution from input audio, approximately 100x higher resolution than conventional Short-Time Fourier Transform (STFT). This signal-processing layer is the foundation of the company's product portfolio, which includes a High-Resolution Spectrogram platform, a Basic Frequency Extraction Solution, a Speaker Verification Solution, and a Speech Recognition Technology, all targeting enterprise customers building voice interfaces for AI speakers, home appliances, automotive systems, and conference call platforms. The company is led by CEO Kim Dong-Jin (김동진) and operates through a wholly-owned US subsidiary, Brainsoft LLC, established in April 2021.

The business model is sales-led and B2B, monetizing the proprietary technology primarily through licensing and royalties rather than disclosed product pricing. Distribution is concentrated in Korean government support channels — including the TIPS program, the Ministry of Science and ICT's AI Voucher Support Project, the AI Data Voucher program from the Korea Data Agency, and TIPS-linked Global Marketing and Commercialization Support — supplemented by a website contact form and the Microsoft for Startups program (selected May 2020). Technical credibility is reinforced by joint research collaborations with the Electronics and Telecommunications Research Institute (ETRI, January 2021) and KAIST (March 2021), and by an IP portfolio that includes one granted Korean patent (registration 10-2164306, October 2020), Korean, PCT, and pending US applications covering DJ Transform and pure-tone extraction methods.

The commercial footprint remains narrow. Total disclosed funding is approximately 300 million KRW (~$220K USD) in angel-linked technology financing surety from Kibo in October 2019, plus an undisclosed angel from POSTECH Holdings (March 2019) and various non-dilutive government grants. No enterprise customers, headcount, ARR, or contract values are publicly disclosed, and pricing is not listed. The company's growth orientation is evident in product launches (January 2020), research partnerships (2021), and US subsidiary formation (April 2021), but commercial scale has not been demonstrated in available sources.

Name

BrainSoft

Legal name

주식회사 브레인소프트 (Brainsoft Inc.)

Website

https://brainsoft.ai

Company type

Private

Founded year

2018

Operating status

Operating

Short description

BrainSoft is a South Korean AI startup that develops high-resolution spectrogram-based voice AI technology using its patented DJ Transform algorithm, licensing speech recognition, speaker verification, and frequency-extraction solutions to enterprise customers building voice interfaces for consumer electronics, automotive, and enterprise communications.

Ownership category

akta.pro rank

Product category

Speech Recognition AI

akta.pro primary industry

Speaker Recognition & Voice Biometrics (HDAAAFAF)

akta.pro secondary industry

On-Device Speech & Audio AI (wake word, ASR, enhancement) (HDAAAJAF)

Keywords

• Speech recognition AI
• Speaker verification AI
• Audio feature extraction
• High-resolution spectrogram
• Voice biometrics

Headquarters

HQ city

Yeoksamdong

HQ country

South Korea

HQ region

Asia

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Operations, Infrastructure

Revenue model

1. AI Speech Recognition Solutions: BrainSoft generates revenue through licensing its high-resolution spectrogram-based AI technology as solutions for speech recognition and speaker recognition. The company develops and commercializes AI solutions targeting AI speakers, home appliances, automotive voice interfaces, and conference call systems.

Go-to-market motion1 record

Distribution channels1 record

Marketing channels3 records

Core offering

BrainSoft develops and commercializes AI voice recognition solutions based on its patented High-Resolution Spectrogram (HRS) technology using the DJ Transform algorithm. The company offers a Basic Frequency Extraction Solution, a Speaker Verification Solution, and a Speech Recognition Technology that deliver approximately 100x higher resolution than conventional STFT-based approaches, targeting AI speakers, home appliances, automotive voice interfaces, and conference call systems.

Product overview

BrainSoft is an AI startup specializing in deep learning and neuroscience-based voice artificial intelligence technology. Their product portfolio centers on High-Resolution Spectrogram (HRS) technology—a core proprietary innovation that extracts audio features at 1Hz/1msec resolution using the DJ Transform algorithm, achieving approximately 100x higher resolution than conventional STFT-based approaches. This core technology serves as the foundation for multiple products: Basic Frequency Extraction Solution (for fundamental frequency analysis), Speaker Verification Solution (for speaker identification), and Speech Recognition Technology. All products work together within the HRS framework, where the high-resolution spectrogram serves as the common pre-processing layer that enables improved performance across speech and speaker recognition applications.

Differentiator

Problem solved

Functional benefit

Products and services

• High-Resolution Spectrogram (HRS) Core proprietary pre-processing technology that extracts spectrograms from sound data at 1Hz frequency resolution and 1msec time resolution — approximately 100x higher resolution than conventional STFT-based methods. Forms the foundational layer that powers BrainSoft's speech recognition and speaker verification products for enterprise voice AI developers.
• Basic Frequency Extraction Solution AI solution that extracts fundamental frequency from voice signals with high precision using HRS technology. Enables accurate voice pitch detection and speaker characteristic analysis, particularly effective for distinguishing speakers by voice characteristics. Targeted at enterprise developers building voice-based biometric and analysis systems.
• Speaker Verification Solution AI-driven speaker recognition solution that leverages high-resolution spectrogram extraction to precisely measure voice pitch, syllable duration, and tone changes. Delivers superior speaker identification accuracy compared to conventional methods, especially in noisy environments, for enterprise customers implementing voice biometrics and authentication.
• Speech Recognition Technology Deep learning-based speech recognition technology that uses high-resolution spectrograms to extract phoneme probability distributions and phoneme time intervals with high precision. Reduces situations where multiple phonemes occur in a single time unit, improving overall speech recognition accuracy for enterprise voice AI products such as AI speakers, home appliances, automotive voice interfaces, and conference call systems.

Quantifiable outcome

• 100x higher resolution compared to conventional STFT (1Hz frequency resolution vs ~32Hz; 1msec time resolution vs ~25msec)
• +1 more outcomes

Segments2 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability4 records

Feature3 records

Partnerships

Three partnerships are on record, tiered core and moderate.

• KAIST (한국과학기술원)coreStrategic or Co-development Partner · 1 March 2021BrainSoft partnered with KAIST in March 2021 for a joint research project on phoneme image identification using high-resolution spectrograms, combining academic research capabilities with BrainSoft's proprietary technology.
• Electronics and Telecommunications Research Institute (ETRI, 한국전자통신연구소)coreStrategic or Co-development Partner · 1 January 2021BrainSoft entered a joint research collaboration with ETRI in January 2021 to conduct research on end-to-end speech recognition using high-resolution spectrogram technology, leveraging ETRI's speech recognition expertise.
• Microsoft for StartupsmoderateGTM or Marketing Partner · 1 May 2020BrainSoft was selected for the Microsoft for Startups program in May 2020, gaining access to Azure cloud resources, Microsoft technology stack, and go-to-market support to accelerate its AI solution development and global expansion.

Scale indicators3 records

Recent moves2 records

Expansion highlights6 records

Direct peers

• Deepgram: Deepgram provides deep-learning speech recognition APIs for enterprise developers, directly comparable to BrainSoft's ASR-focused licensing model. Both target B2B buyers integrating speech AI into products and offer accuracy-focused ASR with emphasis on developer/enterprise channel GTM.
• Speechmatics: Speechmatics offers enterprise speech recognition with strong multi-language and accent/dialect coverage — comparable to BrainSoft's claims around dialect and new-word recognition. Both target enterprise B2B buyers with accuracy-differentiated ASR.
• SoundHound: SoundHound builds voice AI for automotive, IoT, and smart device verticals — directly overlapping BrainSoft's stated target markets (automotive voice interfaces, smart speakers, home appliances). Both pursue licensing-style monetization with voice AI embedded into third-party products.
• Pindrop: Pindrop specializes in voice biometrics, speaker verification, and fraud detection — directly comparable to BrainSoft's Speaker Verification Solution. Both sell into enterprise contact-center and security use cases and compete on accuracy in noisy-channel conditions.
• AssemblyAI: AssemblyAI is a speech-to-text API provider serving enterprise developers with deep-learning ASR models. Highly comparable to BrainSoft in B2B motion, product category (speech recognition), and licensing-style revenue model.

Broad incumbents

• Verint: Verint provides voice biometrics and audio analytics for contact centers and security use cases — overlapping BrainSoft's speaker verification product. A broader incumbent offering wider customer experience analytics alongside voice biometrics.
• Nuance Communications: Nuance (acquired by Microsoft) was the long-standing incumbent in enterprise speech recognition and voice biometrics, serving healthcare, contact center, and automotive verticals. Comparable to BrainSoft in product category and verticals, but at vastly larger scale with an established enterprise channel.
• Google Cloud Speech-to-Text: Google Cloud Speech-to-Text is a hyperscaler ASR service competing across all of BrainSoft's target verticals. A broader incumbent with vastly larger data, compute, and R&D resources — and the most likely hyperscaler to route around DJ Transform via end-to-end architectures.
• iFLYTEK: iFLYTEK is the dominant Asian speech recognition and voice AI provider, with strong presence across automotive, education, and consumer verticals. Comparable to BrainSoft in regional positioning and target verticals, but at vastly larger scale with deep enterprise and government relationships across Asia.

Regional players

• Mindslab: Mindslab is a Korean AI startup focused on conversational and voice AI for enterprise customers. Comparable to BrainSoft as a Korean-domiciled voice AI startup targeting domestic enterprise buyers, though operating with broader conversational AI scope rather than pre-processing-level technology.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat3 records

Key risks5 records

Key highlights6 records

Customer concentration

Revenue estimate

Valuation estimate

Number of profiles

Profiles1 record

Subsidiaries1 record

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about BrainSoft