Skip to main content

Overview

Sully.ai provides built-in SOAP note generation, but many practices need more control over their clinical documentation. When standard SOAP notes do not match your workflow, you have two powerful customization options:
  • Note Styles: Let Sully learn from a sample note you provide
  • Note Templates: Define precise structure with full control over sections, formatting, and content
This guide covers both approaches in depth, helping you choose the right one for your needs.

Note Styles Deep Dive

Note Styles allow Sully to learn your preferred documentation format by analyzing a sample note you provide. This is the fastest way to customize output without defining explicit structure.

How Sully Learns from Your Sample Note

When you create a Note Style, Sully analyzes your sample to understand:
  • Section organization: How you structure information (headers, order, groupings)
  • Writing style: Your preferred tone, sentence structure, and terminology
  • Level of detail: How verbose or concise your notes typically are
  • Formatting conventions: Use of lists, paragraphs, abbreviations
Sully then applies these patterns when generating notes from new transcriptions.

What Makes a Good Sample

Your sample note quality directly impacts the output. Effective samples share these characteristics:

Creating Effective Samples

Follow these guidelines when preparing your sample note:
  1. Include all sections you want: If you want a “Social History” section, include it in your sample
  2. Be explicit about formatting: Use bullet points where you want lists, paragraphs where you want prose
  3. Show preferred terminology: Use your actual clinical abbreviations and phrasing
  4. Demonstrate detail level: If you want thorough assessments, write a thorough assessment

Limitations of Note Styles

While Note Styles are convenient, they have inherent limitations:
  • Less precise than templates: Output may vary slightly between generations
  • Learning interpretation: Sully interprets your sample, which may not always match your intent exactly
  • No structured output: Results are text-only, not JSON with discrete sections
  • Harder to integrate with EHRs: Free-text output requires additional parsing
For use cases requiring consistent, structured output, consider Note Templates instead.

Note Templates Deep Dive

Note Templates give you explicit control over every aspect of your note structure. You define sections, specify prompts for content generation, and set formatting properties.

Full Schema Reference

NoteTemplate

The top-level object that defines your template.

HeadingSection

Defines a section header that can contain nested sections.
Maximum Nesting Depth: Headings can only be nested 2 levels deep. Deeper nesting will cause validation errors.

TextSection

Defines a text content block with AI-generated content. Detail Levels: Tones:

ListSection

Defines a bulleted or numbered list with AI-generated items.

Complete Template Examples

Primary Care SOAP Template

A comprehensive template for general primary care visits.

Cardiology Consultation Template

A specialty template with cardiology-specific sections.

Brief Follow-Up Template

A minimal template for quick follow-up visits.

Using Templates with the API


Alpha: AI Template Generation

Alpha Feature: This endpoint is experimental and may change without notice. See the Alpha API Authentication guide for access details.
Instead of manually building templates, you can use AI to generate them from sample notes or refine existing templates.

Two Modes

Generation Mode

Provide 1-5 sample notes, and Sully will analyze them to create a matching template.

Refinement Mode

Improve an existing template with specific instructions or by showing desired changes.

Authentication Difference

The Alpha API uses Bearer token authentication instead of the X-API-Key header used by the stable API:
For complete API documentation, see:

Migrating from Styles to Templates

When to Migrate

Consider migrating from Note Styles to Templates when you need:

Migration Process

  1. Analyze your style output: Generate several notes using your current style
  2. Identify patterns: Note the consistent sections, headers, and formatting
  3. Map to template sections: Create HeadingSection, TextSection, and ListSection objects
  4. Write effective prompts: Translate your implicit expectations into explicit instructions
  5. Test and iterate: Compare template output to style output and adjust

Example Migration

Original Style Sample (excerpt):
Equivalent Template Sections:

Benefits After Migration

  • Predictable output: Same structure for every note
  • Easier validation: Check that required sections have content
  • Better EHR mapping: Parse JSON directly into your system
  • Fine-grained control: Adjust individual section behavior without affecting others
  • Version control: Store templates as JSON and track changes over time

Next Steps

Note Template Schema

Complete API reference for template objects

Create Note Style

API reference for creating note styles

Alpha: Template Generation

Use AI to generate templates automatically

Notes Guide

Complete guide to note generation