cascade-lattice 0.1.4

Universal AI provenance layer — cryptographic receipts for every LLM call

cascade-lattice

Universal AI provenance layer. Cryptographic receipts for every LLM call.

---

What is CASCADE?

CASCADE creates tamper-proof receipts for AI interactions. Every LLM call—input, output, model, timestamp—gets hashed into a cryptographic chain anchored to IPFS.

One line of code. Every AI call recorded. Forever.

import cascade
cascade.init()

# That's it. Every LLM call now emits a receipt.

---

Installation

pip install cascade-lattice

With provider support:

pip install cascade-lattice[openai]      # OpenAI support
pip install cascade-lattice[anthropic]   # Anthropic support
pip install cascade-lattice[all]         # All providers

---

Quick Start

Option 1: Auto-Patch (Zero Code Changes)

import cascade
cascade.init()

# Now use any LLM library normally
import openai
client = openai.OpenAI()
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello!"}]
)
# ✓ Receipt automatically emitted to lattice

Option 2: Proxy Mode (Any Language)

# Terminal 1: Start the proxy
cascade proxy

# Terminal 2: Point your app at it
export OPENAI_BASE_URL=http://localhost:7777/v1
export ANTHROPIC_BASE_URL=http://localhost:7777/anthropic

# Run any app - Go, Rust, JavaScript, Python, anything
./your_agent
# ✓ All AI calls observed through proxy

Option 3: Manual Observation

from cascade import sdk_observe

sdk_observe(
    model_id="openai/gpt-4",
    input_data="What is 2+2?",
    output_data="4",
    metrics={"tokens": 10}
)

---

How It Works

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│   Your App  │────▶│   CASCADE   │────▶│   OpenAI    │
│             │◀────│   (observe) │◀────│   Anthropic │
└─────────────┘     └──────┬──────┘     │   etc...    │
                           │            └─────────────┘
                           ▼
                    ┌─────────────┐
                    │   Lattice   │
                    │  (receipts) │
                    └──────┬──────┘
                           │
                           ▼
                    ┌─────────────┐
                    │    IPFS     │
                    │ (permanent) │
                    └─────────────┘

Every observation creates a receipt containing:

• Input hash
• Output hash
• Model identifier
• Timestamp
• Merkle proof linking to genesis

Receipts are stored locally in the lattice and can be pinned to IPFS for permanent, decentralized storage.

---

Supported Providers

Provider	Auto-Patch	Proxy
OpenAI	✓	✓
Anthropic	✓	✓
HuggingFace	✓	✓
Ollama	✓	✓
LiteLLM	✓	✓
Cohere	-	✓
Mistral	-	✓
Any OpenAI-compatible	-	✓

---

CLI Commands

# Start proxy server
cascade proxy [--port 7777]

# View lattice stats
cascade stats

# List recent observations
cascade list [--limit 10]

# Pin observation to IPFS
cascade pin <merkle_root>

---

Genesis

CASCADE is anchored to a permanent genesis block on IPFS:

Genesis Root: 89f940c1a4b7aa65
CID: bafkreidixjlzdat7ex72foi6vm3vnskhzguovxj6ondbazrqks7v6ahmei

Every receipt chains back to genesis. The lattice grows. Discovery is reading the chain.

---

Why CASCADE?

• Provenance: Prove what model produced what output
• Auditability: Cryptographic trail of all AI interactions
• Compliance: Meet emerging AI transparency requirements
• Research: Build datasets from cumulative observations
• Trust: Tamper-proof records for AI accountability

---

Links

• Live Demo
• Source Code
• Genesis on IPFS (MessagePack binary - machine-readable)

---

License

MIT License

---

"even still, i grow, and yet, I grow still"