Parts and Themes
When you generate Smart Notes for a lecture-type project, the AI does more than place markers at interesting moments. It performs a structural analysis of the entire transcription, dividing your lecture into logical parts and identifying themes within each part.
A part represents a major section of the lecture — think of it as a chapter. A 60-minute lecture on constitutional law might be divided into three parts: "Historical Context," "Core Principles," and "Modern Applications." Each part groups together a stretch of the recording that covers a coherent subject area.
Within each part, the AI identifies themes — more granular topics that appear in the discussion. Under the "Core Principles" part, themes might include "Separation of Powers," "Due Process," and "Equal Protection." Themes give you a two-level table of contents: parts for broad navigation, themes for precise topic access.
This hierarchical structure is particularly valuable for long recordings. Instead of scrolling through a flat list of dozens of markers, you can expand the part you need, find the relevant theme, and jump directly to that discussion in the audio. The AI creates this structure automatically based on how the content flows — topic transitions, introductory statements, and concluding remarks all serve as signals for where one part ends and another begins.
Content Block Types
Within each theme, the AI generates content blocks — structured pieces of information extracted from the transcript. Rather than simply quoting the transcript verbatim, the AI transforms raw speech into organized, readable formats. There are six types of content blocks, each designed for a specific kind of information:
- Text — summarized paragraphs that capture the main points of a discussion in clear, concise prose
- Quote — exact or near-exact words from the speaker, preserved when the specific phrasing matters
- Table — comparative data, lists of properties, or any information that benefits from a tabular layout
- Logic Chain — step-by-step reasoning, cause-and-effect sequences, or argument progressions shown as connected steps
- Emphasis — key definitions, important warnings, or critical concepts that deserve visual highlighting
- List — enumerated items, steps in a process, or collections of related points
The AI selects the appropriate block type based on the content. A professor listing the five characteristics of a market economy generates a list block. A memorable definition generates an emphasis block. A comparison between two theories generates a table. You can read more about each type in the dedicated Content Block Types guide.
How to Generate Lecture Analysis
Generating a lecture analysis follows the same basic flow as any Smart Notes generation, but the project type matters. To get the lecture-specific analysis with parts, themes, and academic content blocks, your project must be set to the Lecture type. Here is the process:
- Open your project and verify it is set to the Lecture type. If not, change it in Project Settings before generating notes.
- Ensure the recording has a completed transcription. The AI analyzes the transcript text, not the raw audio.
- Tap Generate Smart Notes. The AI begins its structural analysis of the full transcript.
- Wait for processing to complete — typically 1-3 minutes depending on recording length.
- The result appears as a hierarchical structure: parts containing themes, themes containing content blocks, with timestamps linking everything back to the original audio.
The detalization level you choose affects how many parts, themes, and content blocks the AI creates. A Low setting produces a brief overview with fewer blocks. A High setting generates comprehensive coverage with detailed blocks for every significant point in the lecture.
Reading the Results
After generation completes, your lecture analysis appears in the Smart Notes view. The interface presents the hierarchical structure visually — parts at the top level, themes nested within them, and content blocks within each theme.
Each part has a title and a timestamp range showing which portion of the recording it covers. Tapping a part header jumps to its starting point in the audio. Each theme within a part also has its own timestamp, so you can navigate at either level of granularity.
Content blocks appear below their parent theme. A text block shows a concise summary paragraph. A quote block displays the speaker's words in a distinct visual style. A table block renders rows and columns you can scroll through. Each block is connected to a specific moment in the recording — tap it to hear the original audio at that point.
The combined effect is a structured study document that functions as both a summary and a navigation tool. You can read through it sequentially to review the entire lecture, or jump around by topic to find specific information. Use the Smart Notes overview to understand how this fits into the broader AI notes system.
Editing Your Analysis
The AI-generated lecture analysis is fully editable. Nothing is locked — every part title, theme name, content block, and timestamp can be modified to match your needs and terminology.
Common editing workflows include:
- Renaming parts and themes — the AI uses language from the transcript, but you may prefer your own terminology or the textbook's chapter names
- Deleting irrelevant blocks — if the professor went on a tangent about weekend plans, remove that content block
- Adding your own notes — supplement AI-generated blocks with your own observations, questions, or connections to other course material
- Adjusting timestamps — if a block's timestamp is slightly off, drag it to the precise moment you want
- Changing block types — if the AI generated a text summary but you think the content works better as a list, you can change the block type
Editing after generation is where the real value emerges. The AI provides a comprehensive first draft of your lecture notes in minutes. Your job shifts from transcribing and organizing (which the AI handles) to reviewing, refining, and connecting ideas (which requires human judgment). This division of labor can reduce post-lecture note preparation from hours to minutes.