What ai-coustics does

ai-coustics GmbH is a Berlin-based audio intelligence company (founded 2021) that provides a real-time speech enhancement and audio quality SDK for Voice AI systems. The platform unifies five proprietary neural models — Quail (far-field speech enhancement), Quail Voice Focus (near-field speaker isolation), Quail VAD (voice activity detection), Tyto (audio risk scoring), and Rook (perceptual enhancement) — on the proprietary AirTen CPU-based inference engine, which eliminates the need for GPUs, ONNX, or standard ML framework dependencies and executes with 30ms end-to-end latency. Training data spans over 1 million acoustic environments and 500+ noise types, and the system operates language-agnostically across 150+ languages in deployments reaching 187 countries. The product is positioned as an audio reliability layer that improves downstream ASR accuracy (up to 43% word error rate reduction), reduces false barge-ins, and cuts short-utterance failures in production voice agent pipelines.

The company monetizes through a tiered monthly SDK license — Startup at $149/month, Pro at $399/month, Business at $599/month, with usage-based overages — plus custom-priced enterprise contracts for volume deployments. Distribution is product-led via a self-serve developer platform with a 30-day free trial, supplemented by direct enterprise sales, on-premise deployment for data-sensitive customers, and native OEM integrations with LiveKit and Pipecat. Named enterprise customers include PolyAI (2,000+ deployments across 75 languages), Synthesia, Elgato, telli (5M calls), HiDesk, and BosePark Productions. The company has raised approximately $7.4M in disclosed funding across three rounds, most recently a $5.4M Series A in March 2025 led by Partech with participation from Acurio Ventures, arc investors, Connect Ventures, FOV Ventures, and Intuition. Operations are headquartered in Berlin under ai-coustics GmbH (HRB 237856 B), with full GDPR compliance and data residency in the EU.

Name

ai-coustics

Legal name

ai-coustics GmbH

Website

https://ai-coustics.com

Company type

Private

Founded year

2021

Operating status

Operating

Headcount range

11–50 employees

Short description

ai-coustics provides a real-time audio intelligence SDK for Voice AI systems, offering speech enhancement, voice activity detection, and audio risk-scoring models. The Berlin-based company serves voice agent developers, AI avatar platforms, and creator tool companies globally.

Ownership category

akta.pro rank

Product category

Voice AI audio enhancement software

NAICS

Software Publishers (513210)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

Speech Enhancement & Noise Suppression (AEC/NR) (HDAAAFAG)

akta.pro secondary industries

On-Device Speech & Audio AI (wake word, ASR, enhancement) (HDAAAJAF), On-Device Inference Runtimes & SDKs (mobile/embedded) (HDAAAJAB), AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ)

Keywords

• Speech enhancement SDK
• Audio intelligence platform
• Voice activity detection
• Real-time noise suppression
• Voice AI infrastructure

Headquarters

HQ city

Berlin

HQ country

Germany

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Marketing or Sales, Infrastructure, Operations

Revenue model

1. SDK License Subscription: Monthly license subscription with usage-based tiers. Plans include Startup ($149/month), Pro ($399/month), Business ($599/month), and Enterprise (custom pricing). Each plan includes monthly minute allowance, access to all models, and on-premise deployment support.
2. Usage-based Overage: Per-minute pricing for minutes beyond monthly allowance: $0.0015 (Startup), $0.00135 (Pro), $0.0012 (Business). SDK continues to work even when exceeding limits with gentle reminders to upgrade.
3. Enterprise Custom Pricing: Volume-based commercial terms with custom audio evaluations, priority access to new models and beta features, dedicated Slack channel, and volume discounts for large-scale deployments.

Pricing tiers

Go-to-market motion2 records

Distribution channels6 records

Marketing channels9 records

Core offering

ai-coustics sells a real-time audio intelligence SDK that delivers speech enhancement, voice isolation, voice activity detection, and audio quality risk scoring for Voice AI systems. The SDK ships five specialized models (Quail, Quail Voice Focus, Quail VAD, Tyto, Rook) running on the proprietary AirTen CPU inference engine, with native bindings for Python, Rust, Node.js, C, C++, and WebAssembly. It is licensed on monthly subscription tiers with usage-based overages and an enterprise custom-pricing tier.

Product overview

ai-coustics is a real-time audio intelligence platform for Voice AI systems. The product portfolio consists of a unified SDK housing four specialized models: Quail and Quail Voice Focus (speech enhancement for machine processing), Quail VAD (voice activity detection), Tyto (audio quality monitoring and failure prediction), and Rook (perceptual enhancement for human listeners). All models run on the proprietary AirTen CPU-first inference runtime, operate in real-time under 30ms latency, require no GPU or ONNX dependencies, and support 100+ languages. Pricing is subscription-based with per-minute usage tiers from $149/month (Startup) to Enterprise custom quotes.

Differentiator

Problem solved

Functional benefit

Brands

• Quail: Speech enhancement model for far-field and multi-speaker environments, designed to improve STT accuracy for Voice AI.
• Quail Voice Focus
• Quail VAD
• Tyto
• Rook
• AirTen
• Sparrow

Products and services

• Quail Far-field speech enhancement model designed for Voice AI. Improves Speech-to-Text accuracy in challenging environments with background noise, reverb, and multiple speakers by reducing Word Error Rate up to 30%, without suppressing distant-sounding speech, making it suited for speakerphone setups and meeting rooms.
• Quail Voice Focus Near-field primary-speaker isolation model optimized for headsets and handheld devices. Listens briefly at session start before applying suppression, then suppresses competing voices and background noise while preserving the foreground speaker for single-user close-talk voice agent use cases.
• Quail VAD Standalone noise-robust Voice Activity Detection model that predicts speech probability directly from input audio without requiring separate de-noising tools, enabling reliable voice-agent turn-taking, endpointing, and selective audio processing in noisy, multi-speaker environments.
• Tyto Audio intelligence and risk-scoring model that predicts the likelihood of downstream Voice AI failures via a single Tyto Risk Score (0-1) and six qualitative dimensions: noise, speaker reverb, speaker loudness, interfering speech, background media speech, and packet loss. Used for call scoring, quality monitoring, and proactive user interventions in real-time and offline modes.
• Rook Perceptual speech enhancement model that reduces background noise and reverberation while preserving speech naturalness and intelligibility for human listeners. Optimized for real-time constrained systems such as voice calls where natural sound quality matters more than machine-optimized transcription.

Quantifiable outcome

• Up to 43% fewer word errors in noisy environments
• +7 more outcomes

Named customers6 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration2 records

AI capability2 records

Feature9 records

Partnerships

Six partnerships are on record, tiered core and major.

• LiveKitcoreTechnology or IntegrationNative plugin integration allowing real-time speech enhancement in LiveKit voice agents with a single plugin. Compatible with LiveKit versions 0.2.15-0.2.12 using Core SDK 0.17.1.
• PipecatcoreTechnology or IntegrationVoice agent audio pipeline integration with Pipecat. Includes AICFilter for speech enhancement and standalone VAD analyzer. Compatible with Pipecat v1.1.0+ using Python Bindings 2.2.0 and Core SDK 0.17.0.
• PolyAImajorTechnology or IntegrationEnterprise voice agent deployment. PolyAI uses ai-coustics to achieve 40% reduction in false barge-ins and 30% reduction in short-utterance failures across 2,000+ deployments in 75 languages.
• tellimajorTechnology or IntegrationVoice agent scaling partnership. telli scaled to 5 million calls with enterprise-grade reliability using ai-coustics audio reliability layer.
• SynthesiamajorTechnology or IntegrationAI avatar and voice cloning partnership. Synthesia achieves cleaner voice clones, stable speaker identity, and simpler modeling pipeline using ai-coustics for audio preprocessing.
• ElgatomajorTechnology or IntegrationCreator tools partnership. Elgato provides studio-quality sound for millions of creators using ai-coustics SDK running entirely on CPU with VST3 plugin.

Scale indicators12 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Krisp: Krisp provides AI-powered noise cancellation, voice clarity, and meeting assistant SDKs for developers and enterprises. It is the most direct competitor to ai-coustics, operating in the same voice enhancement SDK category with overlapping enterprise customers (BPOs, contact centers, conferencing platforms).
• Adobe Podcast: Adobe Podcast offers AI-powered speech enhancement and noise removal targeted at creators and podcasters, with similar 'Enhance Speech' functionality to ai-coustics' Quail and Rook models. Both compete for the same creator-tools and audio quality use cases.
• Audo.ai: Audo.ai builds AI audio cleaning and noise suppression APIs for developers, targeting voice and video applications with a similar SDK-first developer platform model. It competes with ai-coustics in the speech enhancement API space.

Broad incumbents

• Dolby.io: Dolby.io offers a Media APIs platform that includes noise suppression, voice enhancement, and audio processing for developers. As an established audio brand extending into APIs, it competes with ai-coustics on enterprise-grade speech enhancement with broader media processing capabilities.
• NVIDIA Broadcast: NVIDIA Broadcast and RTX Voice bundle real-time noise removal and voice enhancement into NVIDIA's GPU ecosystem. While not a direct SDK competitor, it provides a hardware-anchored alternative for gamers, streamers, and creators that competes with ai-coustics' VST3 plugin offerings to Elgato customers.
• Deepgram: Deepgram provides ASR and voice AI APIs that increasingly bundle upstream audio preprocessing and noise handling into their end-to-end stack. As an adjacent speech AI platform with overlapping voice agent customers, it represents both a partner and a bundling threat to ai-coustics.
• Microsoft Teams Audio Processing: Microsoft Teams ships integrated noise suppression, echo cancellation, and voice enhancement as part of its collaboration platform. As a bundled incumbent in the conferencing and communication segment, it competes with ai-coustics' RTC and meeting use cases.

Emerging players

• Resemble AI: Resemble AI provides voice cloning and text-to-speech platforms with built-in audio preprocessing and noise robustness. It overlaps with ai-coustics in the voice cloning and AI avatar customer segment (e.g., Synthesia) where clean upstream audio is critical.
• Hume AI: Hume AI builds an emotionally intelligent voice AI platform with proprietary speech understanding models. It is an emerging competitor in the voice agent space that may bundle its own audio handling, potentially reducing demand for ai-coustics' enhancement layer.
• ElevenLabs: ElevenLabs provides AI voice synthesis and voice cloning APIs with growing enterprise voice agent capabilities. While focused on TTS rather than enhancement, it competes in the broader Voice AI infrastructure stack that ai-coustics serves.

Market position

Strengths1 record

Weaknesses1 record

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/ai_coustics
• GitHubgithub.com/ai-coustics
• LinkedInlinkedin.com/company/ai-coustics

Compliance2 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds3 records

Investors7 records

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 ai-coustics