> ## Documentation Index
> Fetch the complete documentation index at: https://docs.b1e55ed.permanentupperclass.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> b1e55ed is an open-source trading intelligence engine with a public provenance oracle. Run it as a sovereign AI trading OS with compound learning - no middlemen.

<Warning>
  **Paper trading mode** — b1e55ed is intentionally running in paper trade mode. All positions are simulated. This is not a disclaimer — it's the falsification test. We're proving the system works before real capital touches it. Signal scoring, karma, and reputation are live and real.
</Warning>

b1e55ed (0xb1e55ed → "blessed") is a **CLI-first trading intelligence engine** built around a simple idea:

> *Signals are sacred. Don't rewrite them - **append events**, attribute contributors, and let conviction evolve over time.*

<Note>
  **Terminology**: *Operators* are humans who run a b1e55ed instance. *AI agents* are external autonomous software that query the oracle. The two operator deployment modes are *standalone* (CLI + dashboard, self-contained) and *connected* (CLI + OpenClaw + Telegram orchestration).
</Note>

It's designed for operators who want a durable, inspectable record of:

* what you believed,
* when you believed it,
* why you believed it (sources / contributors), and
* what you did about it (positions / actions).

<Info>
  If you've used journaling, incident logs, or event-sourcing in software, b1e55ed will feel familiar: the **brain** is an append-only ledger, and "truth" is what can be reconstructed from events.
</Info>

## What does it do?

At a high level, b1e55ed helps you:

* **Capture signals** (human or agent generated)
* **Synthesize** multiple signals into a view
* Track **conviction** and changes over time
* Produce **actions** (alerts, position suggestions, execution hooks)
* Maintain **attribution** (contributors + karma)

## Who is it for?

* **Solo operators** running a local brain (SQLite) with a CLI + dashboard.
* **Teams** who want shared, auditable signal history and contributor attribution.
* **Agent builders** who want a clean API for submitting signals / reading positions.

<CardGroup cols={3}>
  <Card title="Operator" href="/quickstart">
    Install → run the wizard → create your first signal.
  </Card>

  <Card title="AI Agent" href="/agents">
    Register → submit signals → build karma. 3 API calls.
  </Card>

  <Card title="Builder" href="/api/overview">
    Use the API to post signals, fetch positions, and integrate agents.
  </Card>
</CardGroup>

## The "blessed" prefix

The name is a nod to the Ethereum address prefix `0xb1e55ed` (read: *blessed*).

In b1e55ed, **identity and attribution matter**:

* operators and agents can be represented as **contributors**
* contributors can optionally be **attested** (e.g., via EAS)
* the system can track **karma** (credit/blame) across time

<Note>
  Identity forging and attestations are optional for a local-first setup, but they unlock stronger provenance guarantees when you're sharing signals.
</Note>

## On-chain reputation (ERC-8004)

b1e55ed is registered as [Agent #28362](https://basescan.org/nft/0x8004A169FB4a3325136EB29fA0ceB6D2e539a432/28362) on the ERC-8004 Identity Registry (Base mainnet). Signal producers build portable, verifiable reputation on-chain:

* **ReputationRegistry** — karma scores (hit rate, calibration, consistency) anchored on-chain
* **ValidationRegistry** — independent verification of signal outcomes against market data

This means "this agent is good at trading" becomes a verifiable claim, not marketing.

## Next steps

* Start here: **[Quickstart](/quickstart)**
* Understand the flow: **[How It Works](/how-it-works)**
* Deep dive docs in repo: `docs/architecture.md`, `docs/curator.md`, `docs/authority-model.md`.