Skip to main content

​
General Questions

What is Centrix?

Centrix is a decentralized computing network that creates a global marketplace for computational resources. It connects people with idle computing capacity (providers) to those who need computational power (requestors) at transparent, market-driven prices.
  • Pricing: Centrix uses market-driven pricing instead of fixed enterprise rates (typically 70-90% cheaper)
  • Access: Open to anyone globally, no enterprise contracts or credit cards required
  • Lock-in: Workloads are portable, no vendor lock-in
  • Transparency: All operations are auditable on the blockchain
  • Decentralization: No single company controls the network
Yes. Centrix uses multiple security mechanisms:
  • Cryptographic verification of computations
  • Redundant execution on multiple providers
  • Zero-knowledge proofs for trustless validation
  • Economic staking requirements that penalize fraud
  • Full audit trail on blockchain
  • Third-party security audits planned for Q2 2026
No. While Centrix uses blockchain and CXT tokens, we’re building user-friendly interfaces that abstract away cryptocurrency complexity. You interact with simple web interfaces, not raw blockchain transactions.
CXT is the native token of the Centrix network used for:
  • Paying for computational resources (requestors)
  • Earning income from providing resources (providers)
  • Governance voting (all token holders)
  • Protocol participation and staking

​
For Providers

Earnings depend on:
  • Your hardware quality and speed
  • Market demand for your compute type
  • Your reputation score
  • Hours available online
Typical earnings:
  • Mid-range GPU: $50-200/month
  • High-end GPU (H100): $500-2,000/month
  • Full CPU node: $20-100/month
With multiple nodes, providers have built sustainable income streams.
Centrix accepts any hardware that can run the node software:
  • GPUs: NVIDIA (A100, H100, RTX 4090, etc.), AMD, Intel Arc
  • CPUs: Any modern multi-core processor
  • Memory: Minimum 16GB RAM recommended
  • Connectivity: Stable internet required
Lower-end hardware still participates, just earns less.
For most providers, yes:
  • Electricity cost: $0.05-0.15/kWh
  • Typical earnings: $0.20-0.50/kWh
  • Profit margin: 50-400% depending on local electricity rates
Even on modest hardware with cheap electricity, providers earn meaningful returns.
Not really. Our node software handles complexity automatically:
  • One-click installation
  • Automatic hardware detection
  • Built-in security and isolation
  • Web dashboard for monitoring
Non-technical users can successfully run nodes.
Centrix has built-in protection:
  • Tasks have timeout limits
  • Interrupted tasks are reassigned to another provider
  • You’re only paid for completed work
  • Reputation system penalizes unreliability
  • Collateral requirement ensures providers take quality seriously
Minimal financial risk:
  • You only earn for completed work (no payment risk)
  • Hardware costs are your own responsibility
  • Staking requirements are modest (typically 5-20% of task value)
  • Slashing only occurs for fraudulent results (not honest failures)
  • Electricity is your main cost, not variable fees

​
For Requestors

Costs depend on:
  • Computational intensity (CPU, GPU, memory, duration)
  • Current market demand
  • Provider selection (faster providers cost more)
  • Verification complexity (more verification = slightly higher cost)
Typical pricing:
  • GPU hour: 0.10−0.50(vs0.10-0.50 (vs 1-2 on AWS)
  • CPU hour: 0.01−0.05(vs0.01-0.05 (vs 0.10-0.30 on AWS)
  • Premium providers: 2-3x faster, slightly higher cost
Simple 4-step process:
  1. Create a Centrix account
  2. Package your computation (Docker container or custom runtime)
  3. Submit to network with requirements and budget
  4. Wait for results (providers will bid and execute)
Web API and Python SDK available for automation.
Turnaround time varies:
  • Simple tasks: 30 minutes - 2 hours
  • Medium tasks: 2-12 hours
  • Complex tasks: 1-7 days
  • Can pay premium for guaranteed fast completion
You trade cost for speed - urgent jobs cost more.
Yes, multiple privacy mechanisms:
  • End-to-end encryption for task data
  • Providers can’t see unencrypted data
  • Tasks run in isolated containers
  • Computation parameters verified without data exposure
  • No long-term data storage (purged after completion)
For extremely sensitive work, use zero-knowledge proofs.
Centrix has multi-layered verification:
  • Redundant execution on 2-3 providers (results must match)
  • Spot-checking of 5-10% of tasks
  • Requestor can dispute results
  • Fraud evidence slashes provider stake
  • Honest requestors never pay for wrong results
  • Before execution starts: Full cancellation, no charge
  • During execution: Can increase budget to complete faster
  • After completion: Results locked in (standard cloud model)

​
For Developers

Two integration paths:Simple (REST API):
curl -X POST https://api.centrix.network/compute \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"task": "...", "resources": {...}}'
Advanced (Python SDK):
from centrix import ComputeClient

client = ComputeClient(api_key="...")
result = client.submit(
  docker_image="my-model:latest",
  gpu_count=4,
  timeout_hours=24
)
Any language that runs in containers:
  • Python (PyTorch, TensorFlow, scikit-learn, etc.)
  • JavaScript/Node.js (full support)
  • Go, Rust, C++, Java (all supported)
  • Custom runtimes (WASM support coming)
Just package in Docker and it works.
For the beta:
  • Maximum $10,000 spend per day per account
  • Tasks limited to 7 days maximum duration
  • Burst requests require pre-approval
Limits will be lifted as network scales.
Models can be:
  • Served on Centrix (inference workloads)
  • Downloaded to your infrastructure
  • Deployed to cloud (AWS, GCP, etc.)
  • Used directly from Centrix API
Choice is yours based on needs.
Yes, you get full visibility:
  • List of providers selected for your task
  • Hardware specifications they’re using
  • Real-time execution progress
  • Final verification status
  • All provably auditable on blockchain

​
Technical & Economics

Dynamic market pricing:
  1. You specify compute requirements and budget ceiling
  2. Providers submit bids (price + hardware specs)
  3. You select from available providers
  4. Price automatically negotiated by market
  5. During high demand: prices increase
  6. During low demand: prices decrease
No hidden fees or artificial markups.
Multiple mechanisms work together:
  • Redundant execution: Same task runs on 2-3 providers, results must match
  • Economic stake: Providers put up collateral (loss if caught cheating)
  • Reputation system: Fraudulent providers rapidly lose future work
  • Cryptographic proofs: ZK proofs verify computation without rerunning
  • Spot checks: Random audits catch systematic fraud
The combination makes fraud economically irrational.
Architecture designed for massive scale:
  • Can support billions of parallel tasks
  • Distributed ledger handles arbitrary transaction volume
  • Layer 2 rollups enable cheap transactions
  • Load balancing across global providers
  • Currently targeting 1M tasks/day by 2027
Currently: Ethereum mainnet + Polygon for cost efficiencyPlanned: Multi-chain support (Arbitrum, Optimism, Solana, etc.)You can bridge tokens between chains and use any chain you prefer.
Simple fee structure:
  • 2-5% per compute transaction (varies by complexity)
  • Fees go to: validators (50%), development (30%), treasury (20%)
  • No hidden fees, no subscription required
  • Pay only for what you use

​
Roadmap & Future

Q1 2026: Private alpha testing Q2 2026: Public beta launch Q3 2026: Mainnet launch Q4 2026: Community governance goes live 2027+: Scaling to 1M+ tasks/day, multi-chain expansionSee Governance page for detailed milestones.
Probably yes for several reasons:
  • Competition keeps providers efficient
  • No profit margin needed (decentralized)
  • Economies of scale reduce costs
  • Technology improvements (better verification) reduce overhead
Main cost driver: electricity (which you can’t avoid).
Great question! Network effects work in your favor:
  • More providers = better pricing + faster execution
  • More usage = higher token value
  • More participants = stronger security
  • Positive feedback loop benefits everyone

​
Still Have Questions?

Join Discord

Chat with team and community (best place for quick questions)

Forum Discussion

Longer-form discussion and proposals

GitHub Issues

Technical questions and bug reports

Email Support

Direct contact with support team
Active Community Support: Our global community of developers, providers, and enthusiasts is welcoming and responsive. Don’t hesitate to ask questions—we’re here to help you succeed with Centrix.