Open Source AI Protocol for Solana Validators

An Agentic Infrastructure Protocol that makes the Solana Validator experience more efficient.

ESXksGWGPrfCfykwBMDhzwXSAKic8KDmBWhXZhEKpumpCopied!
DocumentationGitHubTwitter

Built a pipeline to automatically spin up Validator

Features

  • Added deploy pipeline for Validator and made it possible to scale if traffic spikes
  • Added data parsing from Prometheus
  • Upgraded dashboard to show all necessary validator details
  • Removed all mock data to let it run on a test validator for better use case

Validator Interaction Mechanism

We finalised how the validators are going to interact with Validate Protocol. We decided on a Server/Client approach since most validators will run on there own servers. So you will be able to run a easy setup process to install the control panel where you then can connect validators to it using the Validate Client. We now have the base done for Validate and can quickly iterate from here.

Features

  • A central Validate Protocol Server hosts the AI agent, protocol core, metrics processor, and Redis store, acting as the coordination and decision making brain
  • Validator Clients run on each validator’s own machine and communicate with the server over gRPC, sending metrics and receiving instructions
  • The Metrics Processor collects validator signals, stores history in Redis, and feeds insights to the AI agent and core for automated or assisted actions
  • The AI Agent, backed by an AI provider, uses the collected data to diagnose issues, recommend fixes, or trigger automated runbooks across connected validators

Validate Agent Initialization

We initialized an Agent that orchestrates a full validator SRE workflow end to end. It continuously collects metrics, evaluates validator health, detects failure conditions, and triggers corrective actions before downtime occurs. The system runs through a Rust-based metrics collector, agent logic, and executor coordinated via Redis, and includes an optional React dashboard that provides real-time visibility into validator status, risk scores, and pending recovery actions.

Features

  • Continuous validator health monitoring with Prometheus style synthetic metrics and rule based issue detection
  • Automated remediation pipeline where the agent identifies issues and the executor performs recovery steps
  • Central Redis datastore used for metrics, action queueing, and historical execution logs
  • Dashboard that provides live validator status, risk scoring, and action queue insights

Initialization

GitHub Account & Repository initialized.