# Thinkn — Belief State Infrastructure for AI Agents

> The `beliefs` SDK gives agents a structured model of what they currently
> believe, why, how strongly, and what is missing. Claims with confidence,
> evidence with provenance, gap awareness, a clarity score (0-1 readiness
> to act), and ranked next moves by expected information gain.
>
> Live docs: https://docs.thinkn.ai/dev
> npm: `beliefs`
>
> Every page below is also available as raw markdown by appending `.md`
> to the URL — fetch those for clean source instead of HTML.

## Start
- [Start Here](https://docs.thinkn.ai/dev/start/index.md): thinkⁿ is world model infrastructure for agents: a durable, auditable belief state that lives outside the model.
- [Install](https://docs.thinkn.ai/dev/start/install.md): Configure your stack and grab the install command + boilerplate. Pick package manager, framework, and memory scope.
- [Quickstart](https://docs.thinkn.ai/dev/start/quickstart.md): The 3-step loop, plus a runnable example that prints clarity rising in under a minute.
- [FAQ](https://docs.thinkn.ai/dev/start/faq.md): Straight answers on what a belief state is, how it differs from memory and RAG, and when to skip it.

## Concepts
- [World model](https://docs.thinkn.ai/dev/core/world.md): The living picture an agent keeps of the world it works in: what's true, how sure it is, and what to do next.
- [Environment](https://docs.thinkn.ai/dev/core/environment.md): The world the agent operates on: the entities and relations that structure it, declared per domain.
- [Observations](https://docs.thinkn.ai/dev/core/observations.md): The input boundary: how evidence enters the world model, and how the agent's own action becomes the next observation.
- [Belief state](https://docs.thinkn.ai/dev/core/beliefs.md): The graph of explicit posteriors over the environment: the belief as the unit, with confidence, evidence weight, edges, and lifecycle.
- [Intent](https://docs.thinkn.ai/dev/core/intent.md): Goals and success criteria, the is/ought firewall, and how moves rank against what the agent is after.
- [Policy](https://docs.thinkn.ai/dev/core/policy.md): The rules of the world: invariants, source hierarchy, conventions, and anti-patterns, and how policy weighs conflicting sources.
- [Actions](https://docs.thinkn.ai/dev/core/actions.md): What the agent can do here: a safety-classified action surface, the predict-then-verify shape, and the act-then-record loop.
- [Worldview](https://docs.thinkn.ai/dev/core/worldview.md): The bounded, decision-sized projection the agent actually reads to act: deterministic, auditable, and capped.
- [Clarity](https://docs.thinkn.ai/dev/core/clarity.md): One score that answers whether your agent has enough to act.
- [Moves](https://docs.thinkn.ai/dev/core/moves.md): VOI-ranked next actions: the engine's answer to where the agent should look next.

## Tutorial
- [Build a Research Agent](https://docs.thinkn.ai/dev/tutorial/research-agent.md): A 30-minute guided build. Learn the model by writing one. End with a working agent that knows what it does not know.
- [Hack Guide](https://docs.thinkn.ai/dev/tutorial/hack-guide.md): Zero to building with beliefs in 10 minutes. Everything you need for the hackathon.

## Reference
- [Core API](https://docs.thinkn.ai/dev/sdk/core-api.md): The beliefs API: what each method does and what comes back.
- [Patterns](https://docs.thinkn.ai/dev/sdk/patterns.md): How to structure your agent loop with beliefs: single-turn, multi-turn, streaming, tool-aware, multi-agent, and the smaller patterns that compose with them.
- [Scope reads](https://docs.thinkn.ai/dev/sdk/reads.md): Plain-English summaries of gaps, decisions, goals, risks, insights, evidence, intents, and contradictions.
- [Moves: SDK](https://docs.thinkn.ai/dev/sdk/moves.md): List, generate, and act on recommended next moves.
- [Trust & tool reliability](https://docs.thinkn.ai/dev/sdk/trust.md): Override agent and source trust at runtime, and track which tools produce useful evidence.
- [Streaming](https://docs.thinkn.ai/dev/sdk/streaming.md): SSE-based live updates for belief state and extraction pipelines.
- [Scoping & Isolation](https://docs.thinkn.ai/dev/sdk/scoping.md): How namespace, writeScope, thread, and contextLayers shape belief state.
- [Auth](https://docs.thinkn.ai/dev/sdk/auth.md): API keys and short-lived scope tokens.
- [Errors](https://docs.thinkn.ai/dev/sdk/errors.md): Every error class the SDK throws, when it fires, and how to handle each one.

## Adapters
- [Claude Agent SDK](https://docs.thinkn.ai/dev/adapters/claude-agent-sdk.md): Use beliefs with the Anthropic Claude Agent SDK. Automatic belief extraction from agent turns.
- [Vercel AI SDK](https://docs.thinkn.ai/dev/adapters/vercel-ai.md): Use beliefs with the Vercel AI SDK. Middleware-based integration for streamText and generateText.
- [React](https://docs.thinkn.ai/dev/adapters/react.md): React hooks for building belief-aware interfaces.
- [DevTools](https://docs.thinkn.ai/dev/adapters/devtools.md): A visual inspector for your agent's belief state.

## Use Cases
- [Enterprise](https://docs.thinkn.ai/dev/cases/enterprise.md): The shared operating state every function's agents read, act, and simulate against.
- [Engineering](https://docs.thinkn.ai/dev/cases/engineering.md): A coding agent that knows what is true before it ships a change.
- [Research](https://docs.thinkn.ai/dev/cases/research.md): Many agents working one body of evidence, their disagreements held open instead of averaged away.
- [Investing](https://docs.thinkn.ai/dev/cases/investing.md): A diligence fleet that gathers evidence to size its conviction in a deal.
- [Legal](https://docs.thinkn.ai/dev/cases/legal.md): An agent that drafts from where the record stands now, not the version three filings ago.
- [Support](https://docs.thinkn.ai/dev/cases/support.md): A support agent that answers from what was promised and shipped, not the thread in front of it.
- [Operations](https://docs.thinkn.ai/dev/cases/operations.md): An ops agent that schedules the right work when supplier backlogs, plant reality, and ERP data keep drifting.
- [Finance](https://docs.thinkn.ai/dev/cases/finance.md): An analyst agent whose thesis tracks the evidence, not the original pitch.
- [Health](https://docs.thinkn.ai/dev/cases/health.md): A clinical agent whose differential stays live until the evidence resolves it.

## Internals
- [Architecture](https://docs.thinkn.ai/dev/internals/architecture.md): Three layers, one loop. How the environment, belief state, and your agent connect, and what happens on every turn.
- [How it works](https://docs.thinkn.ai/dev/internals/how-it-works.md): The lifecycle of a belief: from observation to fused state, with audit and decay along the way.
- [Behavioral contracts](https://docs.thinkn.ai/dev/internals/contracts.md): What the engine guarantees: the behaviors you can build against.

## Optional
- [llms-full.txt](https://docs.thinkn.ai/llms-full.txt): every doc concatenated into one file for single-shot ingestion
- [Sitemap](https://docs.thinkn.ai/sitemap.xml): canonical HTML URLs