Fundraising Fox
Between
Techstars '20Toronto, CA · 1 known investors
Find your way into Between
Sign up to see every warm intro you have to Between
- Paths you didn't know you had: your email and LinkedIn already hold routes to the Between team. We find them
- 2nd- and 3rd-degree connections: the friend-of-a-friend routes you can't see from your own inbox or LinkedIn
- Answers now, not in days: momentum is everything in a raise. Skip asking around whether someone knows someone
Days of research, done the moment you sign in, with every route ranked by how warm it is.
Attention Labs provides selective auditory attention (SAA) technology that enables voice agents to detect whether they are being addressed in multi-party environments and respond only to speech directed at them. The platform offers both cloud and on-device SDKs that integrate with existing voice AI stacks.
Also known as attenlabs · attention labs · Attention Labs
Investors · 1
How we know: the Techstars portfolio · open dataset · Top 3000 Accelerator Startups (no YC) 2026-08 · Not right? Tell us
Company profile
researched Sep 2026Attention labs, operating at between.co, develops an "engagement control layer" for voice AI centred on selective auditory attention (SAA), a real-time addressee-detection system. SAA determines whether speech in a multi-party environment is directed at a device or agent rather than at another person, and runs before audio reaches an automatic speech recognition (ASR) model or large language model. The company positions SAA as distinct from voice activity detection: where VAD detects that someone is speaking, SAA identifies who is being spoken to, forwarding only device-directed speech downstream and suppressing side conversations, background voices and ambient speech. No wake word or hotword trigger is required; the system uses audio and video signals to infer engagement intent.
The product is delivered in two forms: a hosted Cloud SDK with built-in integrations for LiveKit, Pipecat, Twilio and OpenAI Realtime, and an embedded version that runs natively on-device and offline without a cloud round trip. Integration is offered through Python (attenlabs-saa) and JavaScript (@attenlabs/saa-js) client libraries using an API token and a configurable attention threshold. Stated application areas include robotics (a field demonstration with Reachy Mini / OpenRobots in which a robot answers only when addressed), XR and spatial computing with multiple agents in one room performing automatic turn-taking, and personal computers where the voice directed at the device is isolated from other simultaneous speakers.
Business model
The company supplies addressee-detection software to developers and platforms building voice agents, offered as a hosted cloud service and as an embedded on-device component that integrates into existing voice stacks.
Traction
The company cites field demonstrations, including a robotics deployment with Reachy Mini / OpenRobots, references unnamed backers, and was named a CES Picks 2026 winner by TechRadar Pro / TWICE.
Latest developments
The company promotes a CES Picks 2026 award from TechRadar Pro / TWICE and offers both a hosted cloud SDK and an embedded on-device build of SAA.
▸Full profile — market position, technology, go-to-market
Market position
The company frames selective attention, rather than improved wake-word activation, as the requirement for real-world voice agents, and emphasises that SAA operates ahead of speech-to-text, requires no wake word, and can be embedded on-device or consumed via a cloud SDK that plugs into existing voice platforms.
Technology
SAA performs pre-ASR addressee detection in real time, using audio and video signals to infer whether speech is directed at the device. The company reports inference in under 9ms and an end-to-end hosted decision in roughly 150ms, with an audio-visual model reaching 95% F1 in its published evaluation and lighter embedded models around 90%, varying with model size and available signals. The system is designed to fail closed when uncertain. Audio and video are processed transiently on a hosted endpoint, with only device-addressed speech forwarded; raw media is not retained beyond processing or used for model training, and data is encrypted in transit and at rest under a published privacy policy, security page and DPA.
Go-to-market
Self-serve developer adoption via Python and JavaScript SDKs and a free start option, alongside requests for a free evaluation, with prebuilt integrations for LiveKit, Pipecat, Twilio and OpenAI Realtime.
Developers and companies building voice agents for multi-party or noisy environments, including robotics, XR and spatial computing, and personal computing devices.
Compiled by commissioned research from 8 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.
Founder mafia
3 people who came through Between went on to found or lead other companies.
Timeline · 1
launches, deals, and filingsThe company's site cites recognition as a CES Picks 2026 winner from TechRadar Pro / TWICE.
Dated company events from announcements, filings, and press; legal rows summarize public dockets and regulator releases.
▸Research sources · 8
primary sources listed
- Betweenbetween.co · web
8 public sources were cited for this profile; the first-party ones are listed here.
Frequently asked questions
- What does Between do?
- Attention labs builds selective auditory attention software that lets voice agents detect when they are being addressed.
- Who are Between's investors?
- Between's investors include Techstars.
- Where is Between headquartered?
- Between is headquartered in Toronto, CA.