← Back to Home

Documentation

Everything you need to build privacy-preserving AI agents on Solana

Welcome to the KuLM documentation. Here you'll find everything from high-level concepts to low-level implementation details. Whether you're a curious developer exploring the future of autonomous agents, or a battle-tested quant ready to deploy production systems—we've got you covered. Start with the Introduction for the big picture, or jump straight to the SDK if you're ready to build.

CONTENTS

Introduction to KuLM Protocol


KuLM (pronounced "column") represents a paradigm shift in how artificial intelligence operates on public blockchains. In a world where every transaction is visible, where sophisticated actors exploit information asymmetries, and where privacy is often an afterthought—KuLM stands as a beacon of a different future.


The Problem We Solve


Imagine deploying a trading bot on Solana today. Every swap you make is public. Whales can see your positions building. MEV bots front-run your transactions. Your alpha—the edge you spent months developing—becomes worthless the moment it's visible on-chain.


This isn't just a theoretical concern. It's the reality that prevents serious capital from entering DeFi. It's why institutions hesitate. It's why individual traders get eaten alive by sophisticated actors with better infrastructure.


KuLM changes this fundamental dynamic.

What KuLM Provides


KuLM is a zero-knowledge privacy-preserving AI agent protocol built natively on Solana. It enables autonomous agents to:


  • Think Privately: All agent reasoning and decision-making happens in encrypted space
  • Act Confidentially: Execute trades without revealing positions or strategy
  • Prove Integrity: Demonstrate compliance and correctness without exposing data
  • Operate Autonomously: Run 24/7 without human intervention, fully self-sufficient

  • Core Principles


    Privacy First

    Privacy isn't a feature we bolted on—it's the foundation. Every component of KuLM is designed with the assumption that information leakage is unacceptable. We use cutting-edge cryptographic techniques including zk-SNARKs, homomorphic encryption, and secure multi-party computation.


    Trustless Verification

    In decentralized systems, trust is code. KuLM agents can prove they're operating within parameters—capital limits, strategy bounds, compliance rules—without revealing what those parameters are. Verifiers can confirm correctness without accessing secrets.


    Solana Native

    We chose Solana for a reason. 400ms block times enable strategies impossible on slower chains. 65,000 TPS capacity means agent swarms can operate at scale. Sub-cent transaction costs make high-frequency operations viable. KuLM is built to exploit these advantages.


    Open Infrastructure

    KuLM is not a walled garden. It's open infrastructure for a new class of applications. Anyone can deploy agents, build on our SDK, or contribute to the protocol. We believe the future of autonomous intelligence should be permissionless.


    Getting Started


    Ready to deploy your first privacy-preserving agent? Here's the quickstart:


    Install the KuLM CLI

    npm install -g @kulm/cli


    Initialize a new agent project

    kulm init my-agent


    Configure your agent

    cd my-agent && nano kulm.config.yaml


    Deploy to devnet first

    kulm deploy --network devnet


    When ready, deploy to mainnet

    kulm deploy --network mainnet-beta


    Continue to the next section to understand our zero-knowledge architecture in depth.