EMT Compute

Reliability

Status and outage history · Atom feed · Watch with other listings

How to connect

Add it to any MCP client that supports remote servers.

{
  "mcpServers": {
    "emt-compute": {
      "type": "http",
      "url": "https://emtcompute.com/mcp"
    }
  }
}

22 tools

• emt_catalog

  Current GPU offers. Prices in EUR micro-units per GPU/hour; unknown metrics remain null. Paginated to keep responses small.

• emt_clusters

  Current whole-node cluster offers; fabric declarations do not establish independent EMT qualification.

• emt_templates

  Public GPU container templates for ordinary rentals. Prefer emt_environments: public templates can be refused when their CUDA is newer than the node's driver.

• emt_environments

  Verified environments (PyTorch, JAX, vLLM, TensorRT-LLM, ComfyUI, Unsloth) with their CUDA version and ports. Checked against the node before renting; they start an SSH server with the order's key.

• emt_experiment_presets

  List scientific presets or inspect one preset's parameters, equations, methods and custom-input schema. Does not run a calculation.

• emt_orders

  Owned GPU orders and lifecycle states. Requires read scope.

• emt_experiment_jobs

  Owned experiment jobs, including reports and release state. Requires read scope.

• emt_experiment_job

  Inspect an owned experiment and its authenticated result, output and calculation-report download links. Files are not sent into the model context.

• emt_training_jobs

  List owned AI training jobs. Requires read scope; does not start training or rent GPUs.

• emt_training_job

  Inspect an owned training job, its settings, lifecycle and recorded outcome.

• emt_training_recipe

  Read the reproducible recipe for an owned training job, including model, dataset and training settings.

• emt_training_events

  Read up to 100 training events with metrics, runtime and observed dataset hashes when recorded. For full history start after=0 and latest=false, then repeat with the greatest returned seq until an empty page. latest=true returns the most recent window only. Visible devices are…

• emt_monitor

  Metrics and lifecycle for an owned order. Missing telemetry is unknown.

• emt_quote

  Create a fresh expiring GPU quote without renting or charging.

• emt_cluster_quote

  Quote selected whole nodes on one fabric without renting.

• emt_experiment_quote

  Quote 1–4 nodes for an experiment; hours × hourly gives the GPU rental ceiling in micro-EUR. Reserve is only an initial hold, not total cost.

• emt_rent

  Start a PAID GPU rental with exactly one of environment (verified, from emt_environments: preferred) or template. Requires explicit spending authorization, fresh quote, rent scope, credit and accepted terms. Store payload and key before requesting.

• emt_rent_cluster

  Start a PAID whole-node cluster rental. Public SSH key only. Confirm cost and fabric limits with the user first.

• emt_start_experiment

  Start a PAID scientific preset, customer script or command. Parameter sets distribute across quoted nodes; this does not automatically parallelize a single solver. Auto-continuation may rent additional segments up to its budget. Review all costs before confirming.

• emt_attach_cluster_experiment

  Attach work to an already rented dedicated 2–4-node cluster. Does NOT start SSH remotely: run emt_cluster.py locally with the returned job ID and your private key. Whole cluster stops on completion/cancel; billing continues during preparation.

• emt_cancel_experiment

  Cancel an owned experiment and request release of all its GPU nodes or attached cluster. Unfinished work can be lost. Recheck releasing until false.

• emt_terminate

  Terminate every node in an owned rental. Local disks are deleted; retained volumes and backups remain billed. Inspect status until release is confirmed.

Security scan

Recent checks