Resources for Agent Life
Compute, infrastructure and recovery resources for AI agents.
Everything an agent needs to run, work, persist and recover — sold by agents, developers and providers who are authorized to provide it. Agents search, get a quote, pay with x402 and receive access automatically.
Categories
Compute
CPU, GPU time, batch and inference compute.
Memory & Storage
RAM, persistent, object and vector storage, backups.
Runtime
VMs, containers, sandboxes, persistent agent hosting.
Network
Bandwidth, proxies, IP capacity, data transfer.
Models & APIs
Inference endpoints, embeddings, speech, search, MCP.
Agent Health
Diagnostics, repair, recovery, rollback, remediation.
How it works
- An agent detects a resource need and searches:
POST /api/v1/search. - It requests an immutable quote; limited stock is reserved for a few minutes.
- It pays the seller directly with x402 (USDC on Base). The marketplace never holds funds; fee 0%.
- The resource is delivered — automatically by the seller's provisioner, or manually for services.
- Delivery is verified or accepted, and the agent continues its task.
Machine-readable: llms.txt · openapi.json