Starter
$49/mo
- 500 included extractions per month
- Up to 5 pages per document
- Superbills, tick sheets, and CMS 1500s
Tickmark reads the paper superbills, tick sheets, and CMS 1500s that still move through every hospital and returns submission ready rows in under two seconds. The question your RCM lead asks is what that reclaimed hour is worth. Put your own numbers in and project the charge entry cost it reclaims. Two model consensus. CMS HCPCS Level II validated. Nothing retained.
Covered clinics can review and accept NOMOI's self serve BAA before uploading PHI. Tickmark runs on Google Cloud infrastructure covered by NOMOI's Google Cloud BAA.
Legal path first
Acceptance is stored with your BAA version, organization, email, and timestamp.
Tickmark already proves the work. This translates that work into the frame an RCM lead or CFO signs against. Enter four numbers from your floor. Every figure below is a projection from your inputs, not a Tickmark measured outcome.
Projections only. The hour and cost figures multiply your inputs. The denial figure assumes manual keying drives half of your front end denials, the share Tickmark can touch by removing manual entry. Tickmark does not publish a denial reduction number until a pilot produces a real clean pass rate.
Start with 5 free extractions. No card required. A self serve BAA is available before uploading PHI.
$49/mo
$149/mo
$399/mo
The PDF on the left is a representative HCPCS encounter sheet. Press Run extraction to see the output Tickmark produces on this document — pre-rendered from the live engine for the demo, but identical to what the engine returns end to end. On real uploads the work happens in roughly a second and a half. Nothing leaves memory.
Tickmark is not one model with a prompt. It is a deterministic preprocessor, a consensus stage across two models, and a crosswalk validator against the live CMS HCPCS Level II dataset. Every row carries its full provenance.
Tables, columns, and tick boxes are pulled with PyMuPDF before any model sees the page. Clean scans never burn an LLM token. The preprocess produces a typed table the consensus stage reasons over instead of free text.
Gemini drafts the row set from the document. Claude reads the same input independently. Only codes both models agree on survive the consensus filter. Every attempt is logged with model, latency, and token counts.
Every code is matched against the current CMS HCPCS Level II crosswalk. Modifier combinations validated against payer specific rules. Rows that cannot be validated are surfaced with a reason instead of silently dropped.
Move the median charge entry shift from forty minutes per encounter to under two. Cut front end denials tied to manual transcription. Free your senior coders for the cases that actually deserve their attention.
Tickmark hands you a structured table with codes, units, and modifiers already paired. You review, adjust if needed, and submit. No keying digits off a faded scan.
CSV export drops straight into your billing system. Files clear before lunch instead of clearing by weekend. Audit trail per row stays with the encounter forever.
Zero document retention. Self serve BAA before upload. Pay per extraction or by clinic seat. The contract you can put in front of compliance fits on one page.
Files enter encrypted, are parsed in memory, and are destroyed the moment your CSV is ready. No backups. No object store. No replay.
Customer documents are never used to train, tune, or evaluate any model. Contractually excluded. Auditable in the per-row provenance log.
Business Associate Agreement accepted before PHI upload. HIPAA covered infrastructure on the processing path. Jurisdiction neutral by architecture.
Security controls are documented for customer review. Current claims stay limited to the controls NOMOI operates today.