Continuity without extra work

Oncology is long-term medicine, but between two visits there is often an information gap. The first minutes of every consultation are then lost to reconstruction: what has changed since the last visit? LeanOn answers that question before it is asked.

Prepared for every consultation

The prepared overview. Before each visit, LeanOn shows what has happened since the last one: measured trends with treatment context, the patient's self-report, her questions (verbatim, exactly as she asked them) and rule-based alerts. Urgency is decided solely by deterministic, transparently inspectable thresholds (fever, persistent pain, weight loss, silence under active treatment), never by a language model. Every alert names the rule that fired and links to the entries behind it.

Notes that travel. What a doctor pins sits on top at that patient's next visit: "Recheck weight at next cycle." Marking it addressed closes the note; pinning it again carries it one visit further. Clinical notes are immutable: an "edit" creates a superseding note, and the original stays in the record (Section 630f of the German Civil Code, BGB).

Cycles and aftercare in view. Chemotherapy and radiation series are set up in minutes; aftercare plans follow the rhythms of the German S3 guidelines and can be adapted as clinic-wide or personal templates. Overdue follow-up visits appear on the overview by themselves.

The anamnesis draft. A locally running AI writes an interval anamnesis draft: five of its seven sections are rule-built and exact, two are model-written and machine-checked three times over: shape, added value, and every number and date must exist in the source data, or the draft is discarded. It is labelled as an AI draft and never enters the record until a doctor reads, edits and saves it. The signature remains the doctor's.

Your data stays in your clinic

Where the data lives

On your own server, in your own building: no cloud, no external data centre, no LeanOn server. We cannot see your patient data, even if we wanted to. There is no cloud account; patient data is processed and stored on your clinic's hardware. The encrypted access path in online operation runs through Cloudflare (safeguarded by EU Standard Contractual Clauses). Cloudflare sees only connection metadata, never content.

The AI runs in the building

A compact language model runs via local inference on the same server in your building. No API call leaves the clinic. Every AI generation writes an audit entry naming the model.

Who sees what

Role-based, documented and tested: the care team works with the clinical data. Clinic administration has read access to the record (as assistants under medical confidentiality, Section 203 of the German Criminal Code, StGB; in small practices, administration and care team are often the same people); clinical entries can only be made by the care team. LeanOn as the platform operator sees no patient data at all; an automated test checks this on every build. The complete access matrix can be inspected in the product.

What leaves the building

In offline operation: nothing. In online operation, so that patients can connect from home, encrypted tunnel traffic runs through Cloudflare; Cloudflare sees connection metadata, never stored treatment data. And you can switch off public access yourself at any time; the switch sits in your administration area, and in-house access keeps working.

Encryption and accountability

Transport over TLS, documents stored encrypted with AES-256-GCM; two-factor authentication (TOTP) is available for all roles and intended for the care team and administration. Every sensitive action lands in your clinic's audit log: who, what, when, from which IP.

A tour in five screens

Warnings with reasons: every alert arises from a fixed, inspectable rule and links the entries that triggered it. Deterministic, never an AI judgement.

The self-report before the visit: symptoms, medication changes and the patient's own questions reach the care team verbatim.

Notes that travel: what mattered at the last visit lies on top at the next, resolved or carried forward in one gesture.

Aftercare along the S3 guideline: completed, due and upcoming visits across years on one line, instead of spreadsheets and memory.

Accountability: every sensitive action and every AI generation writes a row into your clinic's audit log, including the name of the local model.

A look at the product

Two frames from the running demo environment, synthetic data only.

Physician overview with urgent reports, overdue aftercare and pinned notes — A physician's overview in the morning (German demo clinic).

Pre-visit self-report with the patient's questions, verbatim — A self-report as it arrives (German demo clinic).

See it for yourself

We will show you LeanOn in a personal walkthrough, with synthetic data, and, if you wish, together with your IT team and your data protection officer. The pilot: eight weeks, free of charge, with success criteria defined up front.

Request a demo