Comparing your options honestly
Free is table stakes now. The question is what each option enforces, where your text goes, and who can use it at all.
| Capability | MedSeek (this repo) | Free frontier chatbots (clinician tiers) | Individual scribes | Enterprise scribes |
|---|---|---|---|---|
| Captures audio in the room | No (by design) | No | Yes | Yes |
| Structures notes you already have | Yes, checklist-enforced | Prompt-dependent | Via transcript | Via transcript + EHR |
| Blocks unsafe drafts until gaps have owners | Yes (e.g., pending-results gate) | No | No | Rarely |
| De-identifies before any processing | Yes (assist, not certification) + egress guard | No | Vendors state controls; verify tier | BAA-covered pipelines |
| Citations with retrieval date, or silence | Yes | Varies | N/A | N/A |
| Patient-instruction reading-level scoring | Yes (6th-8th grade target) | No | Sometimes | Sometimes |
| Runs locally / model endpoint is yours | Yes | No | No | No |
| Audit trail of every tool execution | Yes, content-free hash chain | No | Vendor-reported | Vendor-reported |
| Verification wall to sign up | None (install and go) | US NPI + license verification | Email/card trial | Sales process |
| Dated price point | $0, Apache-2.0 | Free (verified US clinicians) [S5][S6] | Freed $39-$119/mo [S1]; Heidi free tier, Clinician $150/mo [S3][S4] | ~$200-$600/clinician/mo, vendor-compiled [S2] |
Prices checked 2026-08-23, re-confirmed 2026-08-26; vendors change tiers — follow the links and verify before deciding. Enterprise figures are a vendor-compiled range, not quotes. Chatbot eligibility: verified US physicians, NPs, PAs, pharmacists per vendor help documentation [S5][S6]. This page describes mechanisms and published prices only; no performance or outcome comparisons are implied. MedSeek price and mechanisms reflect 0.1.3 (nine hosts, 12 lookup tools).
How they compose instead of compete
- Scribes: their output is text. Feed transcripts through MedSeek for structure checks, discharge gating, de-identified teaching copies, and cited lookups.
- Answer engines: keep using them for literature questions. MedSeek's lookup tools exist for label sections and trial records with reproducible queries and retrieval dates attached to the artifact you are drafting.
- Chatbots: anything your policy lets you paste into cloud tools is fine there too. MedSeek covers the cases where policy says no.