All work

NutriOn

Type
Built in-house, application
Made by
Arteca
Languages
Serbian, English
Open the live site(nutrion.rs)
NutriOn website: the headline “Nega koja vidi celu sliku” above a preview of the practice workspace.
The same site on a phone.
Live site, captured 8 November 2026 · nutrion.rs

What it is

Software for nutrition practices. Client records, meal plans and appointments live in one workspace, with an AI assistant that proposes actions and waits for confirmation before saving any change.

What it does

  • Client records, meal plans and appointments in one workspace
  • An AI assistant that proposes actions and waits for confirmation
  • Roles for nutritionists, location managers, owners and reception
  • A separate database and workspace address for each practice

Building around the consultation

A nutritionist needs more than a calorie total to prepare a plan. The client’s assessment, previous consultations, measurements and food preferences all inform the next decision. Reception needs the booking details; the practitioner needs the clinical record.

We built NutriOn to connect that work. A client record brings the assessment and consultation history together. Dated measurements show progress, energy planning records the inputs behind a target, and the calendar connects each visit to its client and nutritionist. The Today view brings upcoming visits, follow-ups and plans awaiting review into each person’s working day.

An assistant that works with the record

The AI assistant opens beside the workspace and understands the page in view. A practitioner can ask for clients due a follow-up, choose an appointment time or prepare a replacement meal plan. For planning, it reads existing plans, recorded dietary preferences and the reason a new plan is needed, then selects foods from the practice’s library.

We made review part of the interaction. Results link to the relevant records, choices appear as selectable options, and proposed changes show a confirmation card before anything is saved. The assistant follows the staff member’s permissions. The nutritionist reviews the proposed foods and suitability for the client before approving a plan.

NutriOn assistant showing an appointment proposal with the client, practitioner, date, duration, and Cancel and Confirm buttons.
An illustrative booking conversation in the real assistant interface, using fictional demo data. The change waits for confirmation.

A meal plan with a history

The editor builds a week from recipes and ingredients, with portions, meal times and alternative meal choices. Energy and macronutrients add up as the practitioner works. Reusable templates give the team a starting point that can be adapted for each client.

Drafts save automatically, while numbered revisions preserve the exact plan reviewed. Approval belongs to a specific revision; later edits are marked as changed since approval. Approved revisions can be exported to PDF where enabled, and staff can record when they handed the plan to the client. That keeps preparation, approval and delivery visible as separate steps.

NutriOn meal editor with a recipe portion, an energy total, a disliked-ingredient warning and an alternative meal containing two foods.
The meal editor with demo recipes and ingredients. A complete alternative meal can contain several foods; the warning flags a recorded dislike.

One practice, several responsibilities

NutriOn models the practice as a team. Owners can work across locations, managers oversee their location, and nutritionists work with assigned clients. Reception can manage contact details and appointments without access to clinical notes, measurements or meal plans. These boundaries are enforced on the server and apply to the assistant too.

Each practice has a dedicated deployment, database and sign-in address. Its locations share a food library within that practice’s instance. This gives the team shared working material while keeping one practice’s records separate from another’s.

Arteca’s part

NutriOn is our own product. We designed and built the practice workspace, its AI workflows and the public marketing website.

The clinic application uses Next.js, React and PostgreSQL; the public site uses Astro. Building both means we work on the product’s daily interactions as well as how it is presented to a new practice. If you’re planning software with complex workflows, permissions or AI assistance, talk to us about your project.

Ask about this project