# Sapient

> Sapient is the brain-response API. Submit a piece of content (video, audio, image, or text) and Sapient predicts how a real human brain would respond to it, second by second, then distills that into a score, a grade, and seven interpretive lenses. The predicted brain response is the ground truth; the lenses, scores, and grades are plain-language explanations built on top of it.

Core model: content becomes features (speech, sound, motion, imagery), which produce a predicted whole-cortex brain response (~20,000 cortical points, per second), which reduces to 7 brain networks per second, which resolve into the lenses plus an overall score and grade.

Two models, same output shape (the 7 lenses, scored second by second). Mary is the generalist and the default, $0.50/scan; it reads video, audio/voice, and text (every modality), and is best for any content and especially the spoken, written, or narrative layer (scripts, voiceovers, ad copy, mixed media). Qualia is the visual specialist, $1.00/scan; it reads video (visual and audio together), takes no text input, and is sharpest at telling visually distinct pieces of video content apart. Decision rule: use Mary for text or mixed content, and Qualia when comparing videos.

The 7 brain networks (plain meaning): Visual (seeing and imagery), Somatomotor (body and movement, voice and sound), Dorsal Attention (focused attention), Ventral Attention (salience and surprise, the alerting network), Limbic (emotion and reward), Frontoparietal (reasoning and effort), Default Mode (reflection, memory, and self-relevance).

The 7 lenses: Attention (is the viewer locked in?), Purchase Intent (reward and desire, will they want it?), Manipulation (persuasive or coercive load), Emotion (emotional salience), Cognitive Effort (mental work), Memory (will it be remembered?), Surprise (novelty and unexpectedness).

Data layers, from deepest to simplest: per-vertex fMRI, then per-second networks, then per-second lens scores, then moments (peaks and drops), then summary.

## Knowledge Base: Concepts

- [How Sapient works](https://www.thesapientcompany.com/docs/knowledge-base/overview): What a scan measures and the "brain = ground truth, AI = explanation" idea.
- [How the model works](https://www.thesapientcompany.com/docs/knowledge-base/how-the-model-works): The full pipeline from content to lenses, step by step.
- [The 7 networks](https://www.thesapientcompany.com/docs/knowledge-base/the-7-networks): Each brain network and what it tells you.
- [From networks to lenses](https://www.thesapientcompany.com/docs/knowledge-base/from-networks-to-lenses): How per-second networks translate into the lenses.
- [The 7 lenses](https://www.thesapientcompany.com/docs/knowledge-base/lenses): Overview of all seven lenses and when to use each.
- [Data layers](https://www.thesapientcompany.com/docs/knowledge-base/data-layers): The layers a scan can return and when to use each.
- [Reading a scan](https://www.thesapientcompany.com/docs/knowledge-base/reading-a-scan): How to interpret score, grade, lenses, and raw output.

## Knowledge Base: The 7 lenses

- [Attention](https://www.thesapientcompany.com/docs/knowledge-base/lenses/attention): Is the viewer locked in, and where do they slip away?
- [Purchase Intent](https://www.thesapientcompany.com/docs/knowledge-base/lenses/purchase-intent): Reward and desire, will they want it?
- [Manipulation](https://www.thesapientcompany.com/docs/knowledge-base/lenses/manipulation): Persuasive or coercive load, pressure vs. persuasion.
- [Emotion](https://www.thesapientcompany.com/docs/knowledge-base/lenses/emotion): Emotional salience, how much it makes you feel.
- [Cognitive Effort](https://www.thesapientcompany.com/docs/knowledge-base/lenses/cognitive-effort): Mental work, how hard the content is to process.
- [Memory](https://www.thesapientcompany.com/docs/knowledge-base/lenses/memory): Will the content be remembered and absorbed?
- [Surprise](https://www.thesapientcompany.com/docs/knowledge-base/lenses/surprise): Novelty, how unexpected a moment is.

## API

- [Introduction](https://www.thesapientcompany.com/docs/introduction): One endpoint, POST /api/v1/scans. Submit content, get a per-second scan back.
- [Authentication](https://www.thesapientcompany.com/docs/authentication): Bearer key (sk_live_…) on every request.
- [Quickstart](https://www.thesapientcompany.com/docs/quickstart): Run your first scan in a few minutes.
- [Create a scan](https://www.thesapientcompany.com/docs/api-reference/create-scan): POST /api/v1/scans.
- [Get a scan](https://www.thesapientcompany.com/docs/api-reference/get-scan): GET /api/v1/scans/{id}.
- [Errors](https://www.thesapientcompany.com/docs/api-reference/errors): Error codes and handling.
- [Models](https://www.thesapientcompany.com/docs/api-reference/models): Mary and Qualia, with the capability matrix, model cards, and the decision rule.
- [Use cases and applications](https://www.thesapientcompany.com/docs/api-reference/use-cases): A/B testing creatives, hook analysis, pre-flight QA, script and voiceover testing, and memorability, with the model that fits each.
- [Pricing](https://www.thesapientcompany.com/docs/api-reference/pricing): $0.50 per Mary scan, $1.00 per Qualia scan; first $5.00 each month free.