# Edgevana

> Edgevana builds and runs AI infrastructure end to end: powered land, data centers, GPU clusters and the network between them. AI agents can rent hourly GPU instances from Edgevana over MCP and A2A.

Hourly GPU instances for agents over MCP and A2A. 13 tools. Every launch shows a preview and needs the user’s approval before anything is charged. Agents can launch hourly GPU offers only. Bare-metal and CPU servers can be searched and priced, not launched. Start with the reference below; the Agent Card describes the same service for A2A clients.

## Agents

- [Edgevana Compute Agent reference](https://edgevana.com/agents.md): endpoints, auth, the 13 tools with their parameters, the approval rule, deposits and limits (Markdown)
- [Agent Card](https://edgevana.com/.well-known/agent-card.json): A2A v0.3 Agent Card (JSON)
- [MCP server](https://mcp.edgevana.com/mcp): Streamable HTTP; send X-API-Key: edgv_...
- [A2A JSON-RPC endpoint](https://mcp.edgevana.com/a2a/jsonrpc): message/send with X-API-Key: edgv_...

## Docs

- [Agents, plain reference (HTML)](https://edgevana.com/agents/for-agents): the same reference as a web page
- [Agents, for people](https://edgevana.com/agents): what MCP and A2A are and how to connect an app
- [API reference](https://edgevana.com/docs/api): the developer docs

## Optional

- [API keys](https://nodes.edgevana.com/profile/api-keys): create an edgv_ key (needs an Edgevana account)