Audio processing location
KerriganOn-premise
AbridgeCloud
FreedCloud
NablaCloud
DeepScribeCloud
AI-powered session transcription and clinical note generation built for behavioral health clinicians — all processing runs on hardware in your building. No patient audio ever reaches an external server.
A 50-minute therapy session can demand 30–45 minutes of documentation. Across a full caseload, that's hours every evening — unpaid, exhausting, and the leading driver of clinician burnout.
Note-taking during a session forces you to divide attention between the client in front of you and the chart you're building. Your clients feel it. Your therapeutic outcomes reflect it.
Every major AI scribe sends session audio to external servers. In behavioral health — where clients share trauma, addiction histories, and legal vulnerabilities — that's a risk your clients never consented to.
⚖ Substance abuse records fall under 42 CFR Part 2 — stricter than HIPAA. External processing creates a disclosure exposure that local-only architecture eliminates entirely.
From audio capture to EHR-ready notes — processed on your Kerrigan device, never in the cloud.
Open the Kerrigan app on your computer before your session. It connects to your practice's Kerrigan device — the on-site backend that handles all processing. Record in-person or connect via Zoom. No push-to-talk, no configuration. Just have a natural conversation.
Whisper-powered transcription runs entirely on your Kerrigan device. Speaker diarization automatically separates clinician and client voices. Nothing leaves your building during processing.
A locally-hosted LLM on your Kerrigan device translates the transcript into a structured clinical note — SOAP, DAP, GIRP, or progress note format. Review, edit, and push to your EHR with one click.
Ask clinically-grounded questions about the session. Surface key themes, risk indicators, or treatment plan updates. Context-aware AI that knows the full transcript — not just a snippet.
Every other AI scribe platform sends session audio to external cloud servers for processing. Kerrigan is different — by design. All transcription and AI inference runs on hardware you control.
Transcription runs on-premise via a locally-hosted Whisper model. No audio uploaded to any external API.
Note generation uses a local LLM. No OpenAI API calls. No Anthropic. No Google. Your PHI never touches their servers.
No patient audio stored in the cloud — ever. Records live only on your hardware and are subject solely to your retention policy.
Fully offline after initial setup. Once provisioned, Kerrigan requires no internet connection. Suitable for high-security environments — substance abuse treatment, forensic settings, court-ordered therapy.
Ships as a managed appliance. We provision and ship a dedicated device to your practice. Plug it in — it configures itself. No IT project, no hardware sourcing, no cloud accounts.
Kerrigan ships as a managed appliance — a dedicated device that sits on your practice network, provisions itself on first boot, and runs all processing locally from that point forward. No cloud accounts, no IT project, no hardware to source.
We ship a preconfigured device to your practice. Connect it to your network and it auto-provisions — installing the Kerrigan backend, AI models, and security configuration automatically. Clinicians install a lightweight app on their computers; everything else is handled for you.
Alternative options for specific environments
Cloud-based competitors process your session audio on external servers. Kerrigan doesn't.
| Kerrigan | Abridge | Freed | Nabla | DeepScribe | |
|---|---|---|---|---|---|
| Audio processing location | On-premise | Cloud | Cloud | Cloud | Cloud |
| LLM inference location | On-premise | Cloud | Cloud | Cloud | Cloud |
| PHI leaves premises | Never | Yes | Yes | Yes | Yes |
| Offline capable | ✓ | ✕ | ✕ | ✕ | ✕ |
| Behavioral health focus | ✓ Core focus | Enterprise | General | General | Oncology |
| Speaker diarization | ✓ | ✓ | ✓ | ✓ | ✓ |
| Zoom integration | ✓ | ✓ | ✕ | ✓ | ✓ |
| HIPAA compliant | ✓ | ✓ | ✓ | ✓ | ✓ |
| 42 CFR Part 2 (substance abuse records) | No external exposure | Cloud exposure | Cloud exposure | Cloud exposure | Cloud exposure |
Session audio captured passively — no push-to-talk, no pauses. Have a natural conversation while Kerrigan listens.
Automatically separates clinician and client voices in the transcript. Your notes reflect who said what.
SOAP, DAP, GIRP, progress notes, and customizable templates — formatted for your specialty and review style.
Direct integration with Zoom for telehealth sessions. Recordings stay local and are never uploaded for processing.
Notes formatted for copy-paste or direct push to major EHR systems. Clean formatting that meets your documentation standards.
Ask questions about the session transcript. Surface risk indicators, treatment themes, or quote exact client language — all grounded in the full session context.
Not a general-purpose scribe retrofitted for therapy — a tool built from the ground up for the narrative, relational, and clinical nuance of behavioral health work.
For substance use treatment & forensic settings: Records governed by 42 CFR Part 2 (SAMHSA) carry disclosure restrictions beyond standard HIPAA. Any cloud-based AI scribe creates a third-party system exposure for those records. Kerrigan's local-only architecture is the only approach that removes that risk entirely.
When clinicians spend less time charting, they see more clients, experience less burnout, and stay longer. Kerrigan compounds those benefits with a data sovereignty posture no cloud tool can match.
Get Early Access for Your Practice →Note on 42 CFR Part 2 & HIPAA: Substance abuse treatment records carry stricter disclosure protections than standard HIPAA. Kerrigan's local-only architecture means those records never enter an external system — eliminating an entire category of compliance risk.
Kerrigan is currently in alpha with individual therapists. We're targeting early April to open onboarding for our first practice cohort — a small group that will shape the product's direction from day one. Waitlist members get priority access, dedicated onboarding support, and founding-member pricing.
No spam. We'll only contact you about early access. You can read our Privacy Policy. Patient data is never collected through this form.