Hanah vs Lyrebird: a plain comparison
We build Hanah, so read this with that in mind. To keep it checkable, everything here about Lyrebird comes from their own public pages, and we link them so you can verify. We hold ourselves to the same standard — we publish our own accuracy and hallucination report, errors included.
What each product is
Lyrebird is an AI scribe built around general practice. Its product is the documentation — consult notes, dictation, letters, and GP care plan paperwork.
Hanah includes the same scribe workflow, and also produces the clinical outputs that follow the note: filled Word and PDF forms, exercise programs with video, outcome measures, and a patient portal with plain-language summaries — all drafted from the same consult.
Side by side
| Hanah | Lyrebird | |
|---|---|---|
| Consult transcription, dictation, and notes | Yes | Yes |
| Note templates | Yes | Yes |
| Works inside your web-based PMS | Browser extension for Cliniko and Halaxy, included in every plan | Best Practice and Gentu, on the Pro plan; free plan states "No EMR Integrations" |
| Document generation | Fills your own Word and PDF templates, GP care plan paperwork included; outputs the DOCX or PDF itself | Letters and documents from templates; GP care plan paperwork |
| Photos and video in the note | Yes — captured in session, flows into notes and documents | No |
| Exercise programs with video | Yes | No |
| Outcome measures (PROMs) | Yes — scheduled send and collection | No |
| Patient portal and plain-language summaries | Yes | No |
| Published error rates | Standing accuracy and hallucination report | Time-saved figures; no vendor-published error rate we could find. An independent hospital trial reports clinician-observed rates |
| Pricing | Free Basic plan (patient portal, exercise programs, PROMs); Practice at $100 per clinician/month ($80 billed annually), everything included, with a 1-month free trial | Free plan metered at 50 consult/dictation actions and 10 documents a month; Pro at A$160 per clinician/month billed annually |
Notes on the table
Working inside your PMS. This one splits by who you are. Lyrebird integrates into general practice software — Best Practice and Gentu in Australia. If you are a GP on Best Practice, that is a genuine advantage over us, because we do not integrate with it. Hanah runs inside Cliniko and Halaxy instead: you open a patient, the panel is already on that patient, and the note files into the note form. Check which plan it sits on either way. As of writing, Lyrebird's free plan states "No EMR Integrations", with integrations on Pro at A$160 per clinician per month. Hanah's is in every plan, including the free one.
Accuracy. We could not find a published error rate or methodology from Lyrebird. The largest independent evaluation we could find is a peer-reviewed trial at Gold Coast Hospital and Health Service — 100 clinicians, 21 specialties, 7,499 consultations. 47% of staff reported hallucinations — output that is factually incorrect, misleading or fabricated, presented as if it were true — with 20% of those saying it happened frequently, and 58% of outputs went into the record without modification.
That study also scored 18 note pairs on a modified PDQI-9 rubric and placed the AI's unedited notes above the clinicians' own, 37.06 out of 40 against 34.6. Taken at face value, the AI wrote the better documents: more thorough, better organised, easier to read. Accuracy was one of the eight domains it won on, and that is the part worth reading closely. There was no independent record of the consultation. A single senior clinician on the project team, with no stated blinding, compared each pair using a transcript the scribe itself had generated — so on the AI's side, the check for accuracy ran partly against the AI's own output. A rubric that rates notes more accurate than a clinician's while half the users report fabrication is not measuring fabrication. Notes that read better are not the same as notes that are true, and neither a quality score nor a transcription percentage is an error rate.
Every ambient scribe hallucinates, ours included. What differs is whether the vendor measures it and shows you how. We publish a standing accuracy and hallucination report with the methodology, the raw machine-readable data behind every number, and our worst results alongside our best. Ask any vendor you evaluate, us included, for their measured error rates and the corpus behind them.
Pricing. Prices move, so check both pricing pages rather than trusting a blog post — including this one. As of writing: Lyrebird's free plan is metered — 50 "actions" a month for consults and dictation, 10 for documents and letters, with no care plans. Unlimited use sits on its Pro plan at A$160 per clinician per month billed annually, or A$240 per clinician per month if you pay monthly. Hanah has a free Basic plan — patient portal, exercise programs, and PROMs, no card required — and its Practice plan, $100 per clinician per month ($80 billed annually) with a 1-month free trial, includes everything in the table — no metered tier, no higher plan.
The bottom line
If all you want is the note, Hanah is one of the best note-writers you can buy — and we publish our benchmarks to prove it. Lyrebird is a strong scribe built around general practice, and it is well integrated into the software general practice runs on.
We are building a different shape of thing. Lyrebird's product is the documentation. Ours is the workflow it sits in — the filled form, the exercise program, the outcome measure and the patient plan, out of the same consult, landing where you already work. If your problem is that the note takes too long, buy the best scribe. If your problem is everything the note is supposed to set in motion, that is what we build.
Either way, run a real trial with real consults before you commit, and hold both of us to our claims.