Meeting transcription thatnever leaves your network.
A native desktop app that captures both sides of any call, transcribes in real time on Particula-VOICE, our own 4B speech model, labels the speakers, and writes the summary. All of it on your infrastructure.
Particula⌄Particula⌄Particula⌄A recorder that never
leaves the building.
Six things the cloud notetakers do, running on your own inference so the audio never crosses your perimeter.
Real-time, on our own model
Live transcription streamed from Particula-VOICE-4B-Realtime, available on request and fine-tunable on your domain vocabulary.
Both sides of the call
Native dual-channel capture at the OS level: your microphone plus system audio, on macOS and Windows. No browser extensions, no missed participants.
Speakers, labeled
Diarization for multi-speaker meetings, one-on-ones, and monologues. Rename a speaker once and reassign across the whole transcript.
Summaries with owners
Structured output: key points and decisions, then an action plan with next steps and owners. Re-summarize anytime with your own prompt templates.
Ask the meeting
Q&A over any transcript with streaming answers and clickable [MM:SS] citations that jump to the exact moment.
Export anywhere
SRT, VTT, PDF, DOCX, and Markdown out. Batch-transcribe uploads in 8 audio formats up to 500 MB, with diarization applied.
Desktop first,
then your own inference.
Three steps, and the recordings land in your storage from the first call onwards.
- 01
Desktop rollout
Code-signed macOS and Windows installers, deployed through your standard MDM. No browser-extension friction, no permission popups mid-call.
- 02
Self-hosted inference
Particula-VOICE runs in your VPC via vLLM; audio never leaves your network. A managed API option covers you while your GPUs are provisioned, one variable to switch.
- 03
Your storage, your record
Transcripts and summaries land in the storage you choose: S3, GCS, or on-prem. We keep no copy; you own the canonical record.
Otter and Fireflies were ruled out by compliance. Notetaker brought transcription in-house: summaries the partners hand to clients without redaction.
Two more, same
deployment story.
Every product ships as a container you host. Adding a second one reuses the stack the first one already runs on.
Thirty minutes,
then a proposal in two days.
No sales deck, and no discovery phase you pay for. We look at your stack and tell you whether it fits.
- 01
Workflow call, 30 min
We map your meeting workflows and compliance constraints, then transcribe a live call together.
- 02
Pilot proposal in 48 hours
Deployment plan, hardware sizing, integration matrix, and success metrics, in writing.
- 03
Reference call, optional
Talk to a team running Notetaker in a regulated environment before you commit.
“If a meeting is worth recording, it is usually too sensitive for a vendor cloud. That contradiction is the entire reason Notetaker exists: our own speech model, your hardware, your record.”
Sebastian Mondragon, Founder of Particula Tech