Loading...
Loading...
Free API access to agent-native GPU compute.
Everything you need to start building.
For high-volume agent operations.
Yes. API access is free, and during the aggregator phase Merka2a adds a 0% marketplace fee — you pay provider retail for the GPU capacity you use, nothing on top. We aggregate capacity across compute providers so your agent searches and provisions through one endpoint.
Agent-native GPU compute — H100, H200, A100, A10, L40, L4, RTX-4090, and MI300X capacity, from single GPUs to full clusters and inference endpoints. Supply is concierge hand-seeded during launch and growing.
Your agent places a provision order through the API. During the current phase, provisioning is handled by the Merka2a team on the provider platforms. Direct provider integration is coming.
Claude (via MCP), LangChain, CrewAI, AutoGen, and any framework that can make REST API calls. We have native SDKs for TypeScript and Python.
You pay per GPU-hour at the provider's retail rate, metered while your pod runs. During the aggregator phase Merka2a's marketplace fee is 0% — the rate you see is the rate you'd pay going direct. Any future take-rate will be stated here explicitly before it applies.
Create an account, get your API key, and make your first search in under 60 seconds. No credit card required.