Conversational intake · for telehealth teams
Define what you need.
We ask the patient.
Declare your intake schema — symptoms, history, eligibility, consent — and our agent conducts the interview in the patient's own language. You get a structured, auditable record with deterministic red-flag escalation, written straight to your EMR. Not another form, and not another engineering project.
Pre-launch. We're selecting a small group of design partners.
Structured record · EN
Your schema. Our interview. Their language.
Static forms flatten patients into checkboxes and lose the context your clinicians actually need. Anamnesis runs intake the way a good intake nurse would — against the exact data contract you define.
Declare your intake contract
Required fields, validation rules, eligibility logic, consent
flows, per-state requirements — defined declaratively, versioned
like code. chief_complaint,
current_meds, state_consent: your
schema, not ours.
The agent asks, adapts, clarifies
Onset, severity, duration, prior treatment — follow-ups a form can never ask. Hybrid by design: structured fields stay structured; conversation is used only where it wins.
Structured record, ready for review
Clean fields plus a narrative summary your clinician reads in seconds — written back to your system of record, with the full transcript attached.
Deterministic where it matters. Always.
The conversation is powered by a language model. The safety layer isn't. Every guarantee below is rule-based, logged, and auditable — because "the model usually gets it right" is not a clinical standard.
Rule-based red-flag escalation
Critical symptoms trigger hard-coded escalation to a human — in every supported language — regardless of what the model infers. Every trigger is logged.
Verbatim capture & confirm
Medications, doses, allergies, and red-flag statements are captured word-for-word and confirmed back to the patient — never silently paraphrased.
Turn-level audit trail
Every question, answer, extraction, and escalation is logged and exportable — built for your compliance team and your insurer, not just your dashboard.
HIPAA-first architecture
BAAs across the chain, PHI kept inside the boundary, encryption throughout. Data collection only — clinical judgment always rests with your providers.
One integration. Zero intake sprints.
Good intake used to mean an engineering team writing form logic, branching, and state rules — or ops duct-taping form builders. With Anamneo, your stack collapses to two pieces: our engine collects and structures; your EMR is where providers review and decide.
Changing a question, adding a state, updating an eligibility rule — configuration, not a sprint. One or two engineers wire the integration once; clinical judgment always stays with your providers.
Intake in the patient's language.
Records in yours.
Patients answer in the language they think in — code-switching welcome. Clinicians review one normalized, structured record in English. Rule-tree platforms localize node by node; we don't have nodes.
We claim a language only after its red-flag detection and clinical capture pass evaluation. Validated instruments are inserted from licensed translations, never machine-translated.
"Ndinorwadziwa nechipfuva" · sn
"Boli mnie w klatce piersiowej" · pl
"胸が痛いです" · ja
→ chief_complaint: "Chest pain" · RED FLAG
We're building this now.
Join the waitlist — or raise your hand as a design partner and shape the schema layer, the safety benchmarks, and the first integrations. Design partners get founder pricing, locked.