Vaultvision MCP server
{
"mcpServers": {
"0xkayser-vaultvision": {
"type": "http",
"url": "https://vaultvision.tech/mcp"
}
}
}{
"mcpServers": {
"0xkayser-vaultvision": {
"type": "http",
"url": "https://vaultvision.tech/mcp"
}
}
}{
"mcpServers": {
"0xkayser-vaultvision": {
"type": "http",
"url": "https://vaultvision.tech/mcp"
}
}
}{
"servers": {
"0xkayser-vaultvision": {
"type": "http",
"url": "https://vaultvision.tech/mcp"
}
}
}{
"mcpServers": {
"0xkayser-vaultvision": {
"type": "http",
"url": "https://vaultvision.tech/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.0.0
What it does
VaultVision exposes read-only research data about Hyperliquid vaults, including search, risk analysis, rankings, and performance metrics. It is for developers building agents or applications that need to analyze vault performance, compare vaults, or monitor risk and alerts.
- Search and retrieve metadata about Hyperliquid vaults to display current APR, returns, and deposit status.
- Compare multiple vaults by risk, entry quality, and performance metrics to help users evaluate options.
- Monitor recent research alerts and risk explanations to track changes in vault conditions.
Tools this server exposes
| Tool | Does |
|---|---|
| search | Search VaultVision's public research surface. |
| fetch | Fetch detailed research data from VaultVision. |
| search_vaults | Search current vault evidence and metadata. |
| get_vault | Retrieve details for a specific vault. |
| rank_vaults | Rank vaults with bounded comparisons. |
| compare_vaults | Compare multiple vaults against each other. |
| explain_vault_risk | Retrieve risk model inputs and caveats for a vault. |
| get_hlp_metrics | Retrieve current HLP evidence and metrics. |
| get_recent_alerts | Fetch retained research alerts about vault changes. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.