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

# Introduction

> Join the Blindference network as a compute provider

# Compute — Introduction

Run a Blindference node to earn fees executing encrypted inference jobs. Help build a private, verifiable, and economically accountable AI execution layer.

## What a Node Does

A Blindference node is a compute provider that:

1. **Attests** its identity to the ICL (Inference Coordination Layer)
2. **Heartbeats** every 60 seconds to prove liveness
3. **Polls** for pending inference jobs every 5 seconds
4. **Decrypts** encrypted prompts via CoFHE threshold FHE under strict ACL
5. **Executes** inference via Groq, Google Gemini, or local vLLM (pluggable backends)
6. **Commits** results back to the ICL for quorum consensus

## Earning Model

Nodes earn fees for successful task completion. Fees are determined by:

* **Task complexity**: More complex prompts earn more
* **Model tier**: Higher-tier models (premium APIs) earn more
* **Attestation tier**: Higher attestation tiers (TPM, TEE) earn premium rates
* **Reputation**: Higher reputation scores get priority assignment

## Attestation Tiers

| Tier | Name    | Hardware            | Trust Level        | Job Eligibility             |
| ---- | ------- | ------------------- | ------------------ | --------------------------- |
| 0    | Mock    | None                | Development only   | Verifier-only, low-value    |
| 1    | TPM 2.0 | TPM chip            | OS-level integrity | Leader + verifier, standard |
| 2    | TEE/SGX | Intel SGX / AMD SEV | Memory encryption  | Leader + verifier, premium  |

**Current status**: Tier 0 (mock) is fully functional for development. TPM and TEE tiers are planned for future releases.

## Requirements

### Minimum (Mock Inference)

* Python 3.10+
* 1 CPU core
* 512MB RAM
* Outbound HTTPS access
* No GPU required

### Recommended (Production)

* Python 3.11+
* 2+ CPU cores
* 2GB RAM
* Stable internet connection
* TPM 2.0 chip (for tier 1)

### Optional (Local Model Inference)

* NVIDIA GPU with CUDA support
* 4GB+ VRAM
* vLLM installed

## Quick Facts

<CardGroup cols={3}>
  <Card title="Setup Time" icon="clock">
    \< 5 minutes with `blindference-node init`
  </Card>

  <Card title="Auto-Healing" icon="heart">
    Auto-re-attests on certificate expiry
  </Card>

  <Card title="Concurrent Jobs" icon="layers">
    Default: 2 concurrent inference tasks
  </Card>
</CardGroup>

## Next Steps

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/compute/quickstart">
    Get your first node running in 5 minutes
  </Card>

  <Card title="Installation" icon="download" href="/compute/installation">
    Detailed installation options
  </Card>
</CardGroup>