Sentient AI Agent Framework Guide

Source: https://sentient.foundation | https://github.com/sentient-agi

Overview

Sentient is an open-source AGI toolstack focused on building Open Monetizable Loyal (OML) AI systems, supporting AI Agent integration with crypto networks.

Installation

pip install sentient-agent-framework

from sentient_agent_framework import SentientAgent

agent = SentientAgent(
    name="my-agent",
    model="openai/gpt-4o",
    system_prompt="You are a helpful Web3 assistant."
)

@agent.tool
def get_token_price(symbol: str) -> dict:
    return {"symbol": symbol, "price": 100.0}

agent.serve(port=8080)

OML-1.0 (Open Monetizable Loyal)

pip install oml-fingerprinting
python -m oml_fingerprinting.embed --model <path> --license oml-1.0

Sentient Enclaves (Confidential AI Compute)

from sentient_enclaves import SecureCompute
compute = SecureCompute(enclave_url="https://enclave.sentient.xyz")
result = compute.run(model="llama-3-8b", input="...", attestation=True)

Fellowship & Grant

• Sentient Fellowship: Supports open-source AGI research, focus areas: OML extensions, Agent toolchains, Web3 AI integration
• Apply: https://sentient.foundation/fellowship
• GitHub: https://github.com/sentient-agi (Sentient-Agent-Framework / OML-1.0-Fingerprinting)

Core Repositories

• Sentient-Agent-Framework: Python Agent framework
• OML-1.0-Fingerprinting: Model copyright protection
• OpenDeepSearch: SOTA search-augmented LLM
• Sentient-Enclaves-Framework: TEE confidential computing

Discord: discord.gg/sentient