Hands-free voice for research labs
We’re building HelixAI to help undergraduate and graduate researchers run protocols, log steps, and set timers using natural speech—tuned for scientific jargon and busy lab environments.
Join the waitlist
Sign up to be notified when HelixAI is ready for pilots in research labs.
Built for the lab, not the living room
Wake‑words per device, jargon‑aware transcription, and campus‑grade admin tools. Start in the cloud, move on‑prem whenever you want.
Device wake‑words
Name each bench: "Helix", "Orion", or custom. Zero‑touch updates from the console.
Jargon aware
Understands ELISA, SDS‑PAGE, qPCR and more, with tenant vocabularies.
Protocol hub
Upload PDFs/DOCX/MD, parse to clean steps, and search by tag, course, or lab.
Campus control
SSO, roles, audit logs, retention policies. FERPA‑friendly defaults.
How it works
Create your college space, connect SSO, and invite lab managers.
Organize by course and semester. Drag‑drop PDFs, we handle parsing.
Plug in bench units, set wake‑words, and start hands‑free experiments.
Simple, campus‑friendly pricing
Start with a free pilot. Scale by device or buy teaching‑lab bundles.
Loved by teaching labs
“My students keep their gloves on and their heads in the protocol. Setup took 10 minutes.”
“Finally a voice system that understands SDS‑PAGE and ELISA without spelling lessons.”
“The wake‑word per bench is brilliant. No more cross‑talk between stations.”
FAQ
Does it work offline/on-prem?＋
Yes. Start in the cloud, or deploy ASR/LLM on your campus or even per-lab. Your data residency, your choice.
Can we set custom wake-words per device?＋
Absolutely. Name each station in the admin console; we push models to devices instantly.
Will it understand our protocols?＋
We ingest your PDFs/DOCX and build a tenant vocabulary so commands map to your steps.
What file types can I upload?＋
PDF, DOCX, and Markdown. We auto-parse headings, steps, timers, and safety notes into a structured protocol.
How does it handle noisy labs and accents?＋
The ASR is tuned for lab noise and supports diverse accents. You can set per-device mic gain and push noise profiles from the console.
Can it set timers and reminders while I’m gloved?＋
Yes. Say “Helix, set a 12-minute spin timer,” or “remind me to invert every 30 seconds.” No phone or keyboard needed.
Does it log what I say back to the protocol?＋
Optionally. You can enable step-by-step voice logs, which attach timestamps, device ID, and user to your run.
Does it integrate with ELNs?＋
Benchling and eLabJournal are on the roadmap. For now, export runs as CSV/JSON or push to webhooks for easy ELN import.
What about privacy and compliance?＋
FERPA-friendly defaults, role-based access, audit logs, and retention policies. PHI/clinical data usage can be restricted or disabled.
Can IT deploy this on campus networks?＋
Yes. Devices use outbound HTTPS only, support WPA2-Enterprise, and can target a campus-hosted API or fully on-prem stack.
How do updates work?＋
Zero-touch. Firmware and models are versioned and staged. You can pin versions per lab or roll back from the console.
Will it trigger other benches by accident?＋
Each device has its own wake-word and acoustic model. Cross-talk filtering and per-bench sensitivity minimize false wakes.
Which languages are supported?＋
English at launch; Spanish and French are coming. You can add tenant terms (genes, reagents) in any language.
What hardware do I need?＋
Our bench unit includes far-field mics and a speaker. You can also run the agent on a small Linux box with a USB mic array.
Is there a student/teaching discount?＋
Yes. Teaching-lab bundles include reduced per-device pricing and priority support during semester start and finals.