Core Feature

Neural Snapshots

Knowledge, not notes.

Every snapshot is a 16-field crystal that captures not just what you think, but why you think it, how to apply it, and how to challenge it. This isn't a highlight — it's a structured piece of how you think.

16 fields. Every angle covered.

Most note-taking tools store words. NeuralSnap stores understanding. Each snapshot captures the full lifecycle of an idea — from where it came from to how to challenge it.

snapshot_card.tsx
nameSOPs Beat Prompts for AI Agents
typeConviction
coreWritten procedures produce 10x better agent output than clever prompts.
proofTested across 6 agents over 1 week. Content quality: unusable → publishable.
confidence0.94
tagsagents, operations, sops

Identity

nametypecoreone_linertags

Origin

storymomentemotionsource

Application

trigger_patternmethodstepsfilter

Challenge

challengebreak_testrisksproof

5 snapshot types

Every snapshot is categorized into one of five types. The type influences how the crystallization pipeline structures the output and helps you filter your brain.

Belief

A subjective conviction about how the world works, shaped by personal experience.

Model

A mental framework or system for understanding a domain — a "how it works" map.

Rule

A prescriptive guideline — something you follow, or think others should.

Conviction

A strongly-held, evidence-backed position you're willing to act on.

Principle

A foundational truth that guides decisions across contexts.

Confidence & significance scoring

Two numeric scores let you quantify how sure you are and how much an idea matters. Both evolve over time as evidence accumulates.

  • Confidence (0–1): How sure are you? 0.5 is a hunch, 0.95 is a bet.
  • Significance (0–1): How much does it matter? Minor observation vs career-defining insight.
  • Break test: What evidence would make you abandon this belief entirely?
  • Challenge: The strongest counter-argument to keep yourself honest.

Search by meaning, not keywords

768-dimensional semantic vectors find knowledge even when the words don't match. Ask 'how to make agents reliable' and find 'SOPs Beat Prompts' — because the meaning is related.

how to make AI agents reliable

3 results · 23ms

SOPs Beat Prompts

Conviction
0.94

Context Window Management

Model
0.87

Document Before Automating

Rule
0.82

Create snapshots with a single API call

Full CRUD via REST API. Create manually or let the crystallization pipeline auto-generate them from raw text.

const snapshot = await neuralsnap.snapshots.create({
  name: "First Principles Thinking",
  type: "Model",
  core: "Break problems down to fundamental truths",
  story: "Elon Musk applied this to reduce rocket costs by 10x",
  confidence: 0.9,
  significance_score: 0.85,
  tags: ["thinking", "problem-solving"],
});

// Search by meaning
const results = await neuralsnap.search(
  "how to approach complex problems"
);
// → "First Principles Thinking" (0.94 similarity)

Built-in challenge framework

NeuralSnap doesn't just store what you believe — it forces you to think about why. Every snapshot includes fields for counter-arguments, break conditions, and risk assessment.

  • challenge: The strongest argument against your belief
  • break_test: What evidence would make you abandon it?
  • risks: What goes wrong if you follow this blindly?
  • proof: Evidence and data that support the belief
  • confidence: Decreases when counter-evidence appears

Use Cases

AI Agent Memory

Store decisions, beliefs, and frameworks your agents can recall semantically. No more re-explaining context every session.

Meeting Insights

Drop a Fathom transcript and get structured snapshots. Six months later, ask "what did we decide about pricing?" and get the answer.

Personal Knowledge Base

Build a searchable, graph-connected library of everything you've learned. Filter by confidence, type, or tags.

Ready to get started?

Create your free account and start building your second brain in minutes.