OpenGradient Docs

Appearance

About OpenGradient ​

OpenGradient is the world's first EVM-compatible blockchain network that extends standard blockchain capabilities with native AI inference, enabling developers to execute ML and LLM models directly from smart contracts with atomic transaction execution. Upload models to the decentralized Model Hub, build on-chain applications with SolidML, deploy off-chain services with the Python SDK, and create personalized AI experiences with MemSync—all with built-in verification through ZKML, TEE, and cryptoeconomic security.

TIP

You can find the OpenGradient block explorer at https://explorer.opengradient.ai.

What is OpenGradient? ​

OpenGradient is an end-to-end decentralized infrastructure network for AI model hosting, secure execution, agentic reasoning, and application deployment.

OpenGradient's novel heterogeneous compute architecture enables permissionless AI inference, statistical analysis, data processing, and agent execution to be securely and scalably run end-to-end from a smart contract. The platform consists of four core components:

  • SolidML: A comprehensive framework for building AI-enabled on-chain applications. SolidML allows developers to securely execute ML and LLM models through simple function calls, with additional capabilities for model scheduling, price feeds, and data preprocessing—all executed as part of atomic transactions.

  • Model Hub: A decentralized model repository built on Walrus storage that allows anyone to permissionlessly upload, browse, and use AI models. All models in the Model Hub are instantly available for inference on the OpenGradient network.

  • Python SDK: A developer toolkit that enables building applications using OpenGradient's decentralized, end-to-end verified AI infrastructure. The SDK provides interfaces for ML and LLM inference, model management, and automated workflows.

  • MemSync: A long-term memory layer for AI built on top of OpenGradient's verifiable inference and embeddings infrastructure. MemSync provides a REST API for long-term context management and AI personalization, enabling developers to build AI applications that remember user context and provide personalized experiences.

OpenGradient powers developers with the ability to seamlessly harness open-source models and leverage compute for features like risk management, ML optimization, AI agents, Sybil resistance, DeFi mechanism design, and more.

OpenGradient Stack

OpenGradient Products ​

  • OpenGradient Network: A highly specialized EVM-compatible network enabling developers to securely and seamlessly host AI models, deploy applications, or execute AI agents. The network supports native AI inference directly from smart contracts with built-in verification through ZKML, TEE, and cryptoeconomic security.

  • SolidML: A Solidity library and framework for building AI-enabled on-chain applications. SolidML provides interfaces and precompiles for ML/LLM inference, data preprocessing, price feeds, and automated model scheduling—all executed atomically within smart contract transactions.

  • Model Hub: A decentralized model repository built on Walrus storage that allows anyone to permissionlessly upload, browse, and use AI models. The Model Hub supports all model architectures from linear regression to neural networks, LLMs, and stable diffusion models. All uploaded models are instantly available for inference on the OpenGradient Network through both smart contracts and the Python SDK.

  • Python SDK: A comprehensive toolkit for building applications with OpenGradient's infrastructure. The SDK enables ML and LLM inference, model management, automated workflows with scheduled execution, and seamless integration with the Model Hub. Available via pip install opengradient with a CLI tool for easy access and iteration.

  • MemSync: A long-term memory layer for AI that provides a REST API for long-term context management and AI personalization. MemSync automatically extracts meaningful memories from conversations, organizes them intelligently, and makes them searchable through semantic search. This enables AI applications to maintain context across sessions and provide truly personalized interactions.

What can you do with OpenGradient? ​

You can build anything with OpenGradient. Here are some examples:

  • On-Chain AI Applications: Upload an AI or ML model to the Model Hub and run inference securely on-chain from a smart contract within seconds using SolidML. All inference is atomically executed as part of EVM transactions with native verification.

  • Dynamic DeFi Applications: Build AMMs with dynamic fee models, lending pools with ML-based risk calculation, or other DeFi protocols that adapt based on model predictions—all running automatically on a schedule.

  • AI Agents: Create LLM-backed AI agents within smart contracts that can make decisions, interact with other contracts, and execute complex workflows based on model outputs.

  • Off-Chain Applications: Build secure web2 AI applications using the Python SDK that leverage OpenGradient's verified and decentralized inference infrastructure, ensuring full integrity and security of AI models and inferences.

  • Automated Workflows: Deploy automated workflows that execute on a schedule with live oracle data, enabling continuous model inference and decision-making without manual intervention.

  • Composable AI Systems: Chain multiple models together using smart contract logic to create composable AI model ensembles or mixture-of-expert applications.

  • Personalized AI Applications: Build AI applications with persistent memory and user personalization using MemSync. Create chatbots, assistants, and context-aware applications that remember user preferences and maintain long-term context across sessions.