Open Playground

zkAPI/PRIVATE ACCESS RAIL

Every request.
A private route.

Put your Sepolia ETHETH to work as private API credits. Reach open AI models and machine services without carrying your payment identity into every request.

Enter the networkOpen playground
  • Fund oncewith Sepolia ETHwith ETH
  • Metered accessacross services
  • Bounded sessionswith zero-knowledge proofs

Public destinations.
Private payment rails.

Fund a private note. Prove your spend. Use any service.
Settle back to a fresh private balance.

  1. 01

    Fund a private note

    Deposit Sepolia ETH into a shared vaultDeposit ETH into a shared vault

  2. 02

    Prove the spend

    Generate a zero-knowledge proof (Groth16) for a bounded session with a one-time nullifier

  3. 03

    Use. Settle. Continue.

    Receive a short-lived, spend-capped key. Usage is metered and settles back to a fresh private balance.

One balance.
An open network.

A single private balance powers all your requests across AI models and machine services.

Fund your balance
Private balance (Sepolia)(Ethereum) Active

0.50 ETH ≈ $ (testnet)

Note lifetime30 days
Provider billing accountsNone
Persistent API keysNone
  • Text Models
  • Image Generation
  • Embeddings
  • Video Generation
  • Ethereum RPC
  • Web Search
  • Agents(coming soon)

Pick the service.
Keep the same rail.

From AI models to machine services.
One private balance. One consistent integration.

View all services
Ready

Text models

Chuteson Bittensor SN64

DeepSeek · Qwen · GLM · Kimi

  • TEE protected inference
  • Decentralized GPU network
Coming soon

Embeddings

Vector representations for AI applications

Coming soon

Video generation

High-quality video generation

Coming soon

Ethereum RPC

Blockchain data and interactions

Coming soon

Web search

Real-time information for your applications

Private access.
Decentralized compute.

ZKRAIL combines a zkAPI-inspired private access layer with Chutes' decentralized inference on Bittensor. Two complementary privacy layers.

Learn about the architecture

Give it a task.
Give it a limit. Coming soon

Autonomous agents with scoped budgets.
Set a spending limit, give it tools, and let it work within your boundary.

Task“Research latest on ZK proofs”
Spending limit$10.00 (testnet)$10.00
Ephemeral spend keyAuto-revokes
  • Web search
  • Ethereum RPC
  • Text models
  • Image generation

The conversation.
Without the billing trail.

Try ZKRAIL in the playground. Chat with open models through Chutes using your private balance.

  • Browser-local history (optional)
  • No chat history on server
  • Uses a short-lived, spend-capped key
  • Metered usage with settlement
Open playground

Proof carries value.
Identity stays apart.

ZK proofs authorize usage without revealing your identity. Each request is backed by cryptographic primitives.

Learn about zk proofs
  1. 01

    Commit

    Create a commitment (Poseidon, 32-level Merkle tree)

  2. 02

    Authorize

    Generate a Groth16 proof on BN254 for a bounded session with a one-time nullifier

  3. 03

    Settle

    Usage is metered and settles back to a fresh private balance

A simple architecture
for a more open future.

ZKRAIL routes private access to open AI models and machine services, powered by zkAPI-style payments and Chutes on Bittensor.

Your ApplicationAgents · Chat · IDE · Services
ZKRAILSession router & orchestration
Private accesszkAPI-style layerPrivate balance · ZK proofsCapped keys · Usage receipt
Open AI computeChutes · decentralized inferenceTEE confidential computeBittensor SN64 operators
Testnet
◇Ethereum SepoliaEthereum MainnetPrivate balance
τBittensor SN64Compute network
Mainnet models viaOpenRouterOpenAI

Your next request.
On your terms.

Sepolia testnetEthereum mainnet · Experimental, unaudited protocol.
Use testnet ETH only.Uses real ETH. Start with a small amount.

Open the playgroundExplore the network