Stop keying W-2s
into Drake.
AutoKey reads your client’s source documents, extracts every value, and shows you each field side by side with the PDF. Nothing pushes to Drake until you approve it.
Built for your client data.
Deterministic, not AI vibing
We extract every value, you approve every field before anything pushes to Drake. No surprise hallucinations in client returns.
Zero data retention
Source documents and extracted values are deleted after each session. We never train on your client data.
17 forms live
W-2, every 1099 variant, K-1 (1065/1120S/1041), 1098, 1098-T, SSA-1099. Drake Tax 2025.
Security on every return, not an afterthought.
Data Encryption
Every uploaded document and extracted value is encrypted with AES-256 at rest and TLS 1.3 in transit.
Infrastructure
Hosted on Supabase Cloud (SOC 2 Type 2, ISO 27001 certified) in US-based data centers, with DDoS protection included.
Access Controls
Role-based permissions and multi-factor authentication (MFA), with granular team-member access scopes for every client.
Audit Trails
Every upload, extraction, review, and push to Drake is logged with actor, timestamp, and IP, retained up to 1 year and exportable on request.
Compliance
Built on SOC 2 Type 2 certified infrastructure. Actively pursuing our own SOC 2 Type 1 certification. GDPR compliant.
AI Data Privacy
Zero Data Retention (ZDR) on all AI processing: your client documents are never stored by AI services or used for model training. Context is discarded the moment a return is extracted.
Not on Drake?
ProConnect, CCH Axcess, Lacerte, and UltraTax are next. Email me which one you need and I’ll add you to the list.
A 200-return season is
130 hours of typing.
Every W-2, 1099, and K-1 gets typed into Drake by hand. Number by number, form by form, client by client. That’s hours you can’t bill, can’t review, and can’t get back.
That’s the work AutoKey takes off your desk.
Four steps.
You approve before anything pushes.
Upload the PDFs
Drop in your client's W-2s, 1099s, K-1s, 1098s. PDFs, scans, or photos. All accepted.
AI extracts every field
AutoKey reads each document and pulls out every value: Box 1, EIN, payer name, interest paid, whatever the form carries.
Review side by side before anything commits
Document on the left, extracted values on the right. Confirm or fix each field. Nothing moves into Drake until you approve it.
AutoKey pushes to Drake
Approved values land on the correct Drake screen and the correct line. Same mapping, same result, every time.
Every source document
your clients send.
17 forms live in Drake Tax 2025. New ones ship every week. If yours isn’t here, request it.
Pay per return.
Cheaper as you scale.
No subscription and no per-seat fees. You only pay for the returns you run.
Try it free. Every new account gets 5 returns to test on your own client documents. No credit card.
Pay per return
from $12.50Volume discounts, billed by the return.
- 1return$16/ return
- 50returns$15.50/ return
- 100returns$15/ return
- 200returns$14/ return
- 300returns$13/ return
- 500returnsBest rate$12.50/ return
Volume is cumulative. Buy 100 returns now and 400 later, and you’re billed at the 500-return rate, not as two separate tiers.
Enterprise
CustomFor high-volume & multi-office firms.
- Custom rate below $12.50 per return
- Support beyond Drake: ProConnect, CCH Axcess, Lacerte, UltraTax
- Unlimited preparers & seats
- SSO / SAML
- Advanced audit logs & exports
- Uptime SLA & priority support
- Dedicated onboarding & account manager
Plain answers.
Anything not covered here, drop it in the headache field on the form or email contact@taxhance.com.
Which versions of Drake does AutoKey work with?
What happens when the OCR misreads a number?
How does it handle K-1s, multi-state returns, and complex entities?
Do you use AI on my client data?
What if I'm not on Drake?
How is this different from other AI tax tools?
Three things:
- 1Reviewable. Every extracted value is shown side by side with the source document before anything pushes to Drake, so a human is always the last one to approve.
- 2Zero data retention. Your source documents and the values we extract are deleted the moment your session ends. They are never stored long-term.
- 3No model training. Your client data is never used to train our models or any third party model.