For Requestors

​

Getting Started

​

Create Account

1. Visit centrix.network
2. Sign up with email or Web3 wallet
3. Verify email address
4. Set up payment method (CXT wallet or credit card bridge)

​

Fund Your Account

# Option 1: Direct CXT transfer
centrix wallet transfer \
  --to centrix-account:user@example.com \
  --amount 100 CXT

# Option 2: Swap stablecoins for CXT
# Via Uniswap, Curve, or other DEXs

# Option 3: Credit card onramp
# Automatic conversion to CXT via partner service

​

Create Your First Task

# task.yml
name: "Blender Render - My Scene"
description: "Render 100 frames of animated scene"

requirements:
  cpu_cores: 8
  memory_mb: 16000
  storage_gb: 100
  gpu: "NVIDIA RTX 3080+"
  duration_estimate: "2 hours"

input:
  files:
    - source: "s3://my-bucket/scene.blend"
      destination: "/input/scene.blend"
    - source: "local:textures/"
      destination: "/input/textures/"

execution:
  container: "centrix/blender:4.0"
  command: [
    "blender", "-b", "/input/scene.blend",
    "-o", "/output/frame_####",
    "-f", "1..100"
  ]

output:
  files:
    - source: "/output/frame_*.png"
      destination: "s3://my-bucket/output/"

verification:
  method: "redundant"
  redundancy_factor: 3
  threshold: 0.67

pricing:
  max_price: 50 CXT
  urgency: "normal"

​

Submit and Monitor

# Submit task
centrix task submit task.yml

# Monitor progress
centrix task status <task-id>

# View real-time logs
centrix task logs <task-id> --follow

# Download results when complete
centrix task download <task-id> --output ./results/

​

Task Types and Templates

​

Pre-built Templates (Quick Start)

• 3D model rendering
• Animation sequences
• Batch rendering
• Pre-configured environment

• PyTorch/TensorFlow support
• Distributed training setup
• GPU acceleration
• Custom dataset handling

• FFmpeg transcoding
• NVIDIA GPU acceleration
• Batch processing
• Multi-format support

• Python/NumPy/SciPy
• MATLAB compatibility
• Large dataset processing
• Distributed algorithms

• SQL queries on large datasets
• Statistical analysis
• ETL pipelines
• Custom Python scripts

​

Custom Task Development

# Create custom task from Docker image
centrix task create-custom \
  --image "my-organization/my-app:latest" \
  --name "custom-processing" \
  --cpu 8 \
  --memory 16000

# Test locally before submitting
centrix task test ./my-task.yml --local

​

Cost Optimization Strategies

​

1. Off-Peak Scheduling

• Submit during low-demand periods (nights, weekends)
• Automatic scheduling available
• Typical savings: $10-$50 per task

​

2. Batch Processing

• Combine multiple small tasks into larger batches
• Better provider selection and negotiation
• Lower per-task coordination costs

​

3. Smart Provider Selection

• New providers offer 20-40% discounts
• Geographic proximity reduces latency costs
• Reputation system protects quality
• Balance cost vs. performance requirements

​

4. Redundancy Optimization

• Critical tasks: Full redundancy (3x verification)
• Standard tasks: Spot verification (30% random checks)
• Low-risk tasks: Single execution with hash verification
• Savings: Up to 60% on verification costs

​

Real Pricing Comparison

​

Task Management Features

​

Real-Time Monitoring

• View provider details and reputation scores
• Track resource utilization (CPU, memory, network)
• Monitor task progress and estimated completion
• Receive alerts on issues or completion

​

Multi-Task Management

• Batch submit related tasks
• Priority queuing for urgent work
• Historical task tracking and analytics
• Cost breakdowns and budget management

​

Issue Resolution

• Automatic retries on provider failure
• Seamless provider reassignment
• Built-in dispute resolution system
• Insurance pool for task recovery
• 24/7 support response

​

Getting Support