zeroproof 0.2.4

Python SDK for ZeroProof AI verification API - Secure your agentic e-commerce ecosystem

ZeroProof Python SDK

Python SDK for the ZeroProof AI verification API. Secure your agentic e-commerce ecosystem with zero-knowledge proofs.

Installation

pip install zeroproof

Quick Start

from zeroproof import ZeroProof

# Initialize the client with your API key
client = ZeroProof(api_key="zkp_your_api_key_here")

# Create a verification challenge
challenge = client.create_challenge(
    agent_id="shopping-assistant-v1",
    action="add_to_cart",
    context={"item_id": "laptop-123", "price": 999.99}
)

print(f"Challenge ID: {challenge.challenge_id}")
print(f"Nonce: {challenge.nonce}")
print(f"Expires in: {challenge.expires_in} seconds")

# Verify the proof (in production, this would be generated by your agent)
result = client.verify_proof(
    challenge_id=challenge.challenge_id,
    proof="your_cryptographic_proof_here",
    agent_signature="optional_signature_here"
)

if result.verified:
    print(f"✅ Verification successful!")
    print(f"   Agent: {result.agent_id}")
    print(f"   Action: {result.action}")
    print(f"   Confidence: {result.confidence * 100}%")
    # Now execute the actual e-commerce action
else:
    print("❌ Verification failed")

Features

• 🔐 Zero-Knowledge Proof Verification: Verify agent authenticity without revealing sensitive data
• 🔒 Encrypted Messaging: Secure agent-to-agent communication with AES-256-GCM
• 🚀 Simple API: Easy-to-use Python interface
• ⚡ Fast: Low-latency operations (< 100ms for encryption)
• 🛡️ Secure: End-to-end encryption with automatic key management

Use Cases

Encrypted Agent Communication

from zeroproof import ZeroProof

client = ZeroProof(api_key="zkp_...")

# Agent 1 sends encrypted message
result = client.send_encrypted(
    to_agent_id="checkout-agent",
    message={
        "action": "process_order",
        "order_id": "ORD-12345",
        "amount": 299.99
    },
    ttl_minutes=60
)

print(f"Message ID: {result.message_id}")

# Agent 2 receives and decrypts
message = client.receive_encrypted(message_id=result.message_id)
print(f"Order: {message.message['order_id']}")
print(f"Amount: ${message.message['amount']}")

E-Commerce Agent Authorization

from zeroproof import ZeroProof

client = ZeroProof(api_key="zkp_...")

# Before allowing an agent to make a purchase
challenge = client.create_challenge(
    agent_id="checkout-agent-v2",
    action="initiate_purchase",
    context={
        "amount": 1499.99,
        "merchant": "TechStore",
        "items": ["laptop", "mouse", "keyboard"]
    }
)

# Agent generates proof
proof = your_agent.generate_proof(challenge)

# Verify before processing payment
result = client.verify_proof(challenge.challenge_id, proof)

if result.verified and result.confidence > 0.95:
    process_payment(1499.99)
else:
    log_security_incident(result)

Refund Authorization

# Verify agent before approving refund
challenge = client.create_challenge(
    agent_id="refund-processor",
    action="approve_refund",
    context={"order_id": "ORD-12345", "amount": 299.99}
)

result = client.verify_proof(challenge.challenge_id, proof)

if result.verified:
    approve_refund(order_id="ORD-12345")

Inventory Management

# Verify agent before updating inventory
challenge = client.create_challenge(
    agent_id="inventory-manager",
    action="update_stock",
    context={"product_id": "PROD-789", "quantity_change": -50}
)

result = client.verify_proof(challenge.challenge_id, proof)

if result.verified:
    update_inventory("PROD-789", quantity=-50)

API Reference

ZeroProof

The main client class for interacting with the API.

__init__(api_key: str, base_url: Optional[str] = None)

Initialize the ZeroProof client.

Parameters:

• api_key (str): Your ZeroProof API key (starts with 'zkp_')
• base_url (str, optional): Custom API base URL (defaults to production)

Raises:

• ValueError: If API key is invalid or missing

create_challenge(agent_id: str, action: str, context: Optional[Dict] = None) -> Challenge

Create a new verification challenge.

Parameters:

• agent_id (str): Unique identifier for the AI agent
• action (str): The action the agent wants to perform
• context (dict, optional): Additional context for the action

Returns:

• Challenge: Object containing challenge_id, nonce, expires_in, timestamp

Raises:

• ZeroProofError: If the request fails

verify_proof(challenge_id: str, proof: str, agent_signature: Optional[str] = None) -> VerificationResult

Verify a proof for a given challenge.

Parameters:

• challenge_id (str): The challenge ID from create_challenge()
• proof (str): The cryptographic proof data
• agent_signature (str, optional): Optional agent signature

Returns:

• VerificationResult: Object containing verification details

Raises:

• ZeroProofError: If verification fails

get_status(session_id: str) -> Dict

Get the status of a verification session.

Parameters:

• session_id (str): The session/challenge ID to check

Returns:

• dict: Session status details

Raises:

• ZeroProofError: If the request fails

Encrypted Messaging

send_encrypted(to_agent_id: str, message: Any, ttl_minutes: int = 60) -> EncryptedMessage

Send an encrypted message to another agent.

Parameters:

• to_agent_id (str): Target agent identifier
• message (Any): Message content (string, dict, or JSON-serializable data)
• ttl_minutes (int, optional): Time-to-live in minutes (default: 60, max: 1440)

Returns:

• EncryptedMessage: Object containing message_id, expires_at, status, ttl_minutes

Example:

result = client.send_encrypted(
    to_agent_id="agent_456",
    message="Hello, secure world!",
    ttl_minutes=30
)

receive_encrypted(message_id: str) -> DecryptedMessage

Receive and decrypt an encrypted message.

Parameters:

• message_id (str): The message ID from send_encrypted()

Returns:

• DecryptedMessage: Object with decrypted message and metadata

Raises:

• ZeroProofError: If message not found, expired, or decryption fails

Example:

message = client.receive_encrypted(message_id="msg_abc123...")
print(f"Content: {message.message}")
print(f"Read count: {message.read_count}")

Data Classes

Challenge

Represents a verification challenge.

Attributes:

• challenge_id (str): Unique challenge identifier
• nonce (str): Random nonce for this challenge
• expires_in (int): Time until expiration in seconds
• timestamp (int): Challenge creation timestamp

VerificationResult

Represents the result of proof verification.

Attributes:

• verified (bool): Whether the proof was verified
• agent_id (str): The agent that was verified
• action (str): The action that was authorized
• confidence (float): Confidence score (0.0 to 1.0)
• timestamp (str): Verification timestamp
• session_id (str): Session identifier

EncryptedMessage

Represents an encrypted message response.

Attributes:

• message_id (str): Unique message identifier
• expires_at (str): Expiration timestamp (ISO 8601)
• status (str): Message status ("ready")
• ttl_minutes (int): Time-to-live in minutes

DecryptedMessage

Represents a decrypted message.

Attributes:

• message_id (str): Message identifier
• from_agent_id (str): Sender's API key (partial)
• to_agent_id (str): Recipient identifier
• message (Any): Decrypted message content
• read_count (int): Number of times message was read
• created_at (str): Creation timestamp
• expires_at (str): Expiration timestamp

Exceptions

ZeroProofError

Base exception for SDK errors.

Attributes:

• message (str): Error message
• status_code (int, optional): HTTP status code
• response (dict, optional): Full API response

Context Manager Support

The SDK supports context managers for automatic resource cleanup:

with ZeroProof(api_key="zkp_...") as client:
    challenge = client.create_challenge("agent-1", "purchase")
    result = client.verify_proof(challenge.challenge_id, proof)
# Session is automatically closed

Error Handling

from zeroproof import ZeroProof, ZeroProofError

client = ZeroProof(api_key="zkp_...")

try:
    challenge = client.create_challenge("agent", "action")
    result = client.verify_proof(challenge.challenge_id, "invalid_proof")
except ZeroProofError as e:
    print(f"Error: {e.message}")
    print(f"Status Code: {e.status_code}")
    print(f"Response: {e.response}")

Handling Message Expiration

try:
    message = client.receive_encrypted(message_id="msg_...")
except ZeroProofError as e:
    if e.status_code == 410:
        print("Message expired")
    elif e.status_code == 404:
        print("Message not found")
    else:
        print(f"Error: {e.message}")

Configuration

Custom API Endpoint

For testing or self-hosted deployments:

client = ZeroProof(
    api_key="zkp_...",
    base_url="https://api.custom-domain.com/v1"
)

Environment Variables

You can set your API key via environment variable:

import os
from zeroproof import ZeroProof

api_key = os.getenv("ZEROPROOF_API_KEY")
client = ZeroProof(api_key=api_key)

Requirements

• Python 3.8+
• requests >= 2.25.0

Examples

See the examples directory for complete examples:

• shopping_agent_demo.py - ZKP verification demo
• encrypted_messaging_demo.py - Encrypted messaging demo with multiple scenarios

Run an example:

export ZEROPROOF_API_KEY="zkp_your_key_here"
python examples/encrypted_messaging_demo.py

Getting an API Key

1. Sign up at https://zeroproofai.com
2. Navigate to your dashboard
3. Generate a new API key
4. Copy the key (starts with zkp_)

Development

Installing for Development

git clone https://github.com/jacobweiss2305/zeroproof.git
cd zeroproof/python-sdk
pip install -e ".[dev]"

Running Tests

pytest tests/

Code Formatting

black zeroproof/
flake8 zeroproof/
mypy zeroproof/

Support

• Documentation: https://docs.zeroproofai.com
• Issues: GitHub Issues
• Email: [e-mail]

License

MIT License - see LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Changelog

0.2.0 (2025-10-02)

• ✨ Added encrypted messaging service
• 🔒 AES-256-GCM encryption for agent-to-agent communication
• ⏰ TTL-based message expiration with auto-cleanup
• 📖 Multiple reads supported until expiration
• 📝 New example: encrypted_messaging_demo.py

0.1.0 (2025-01-10)

• Initial release
• Basic challenge/proof verification flow
• Context manager support
• Comprehensive error handling
• Full type hints