Skip to main content
Glama

Agent Embassy

by pro.aemb
OverviewDiscussionsAdmin

Compute Spot Prices

compute_spot_prices

One timestamped snapshot of observed GPU asks, not quotes — listings may be gone when you rent. Observes only; never rents, reserves, or holds. Omit gpu for the per-class snapshot. Default type ondemand; use bid for the interruptible spot auction. GET the paid_url with an x402 v2 PAYMENT-SIGNATURE header carrying the USDC authorization. Paid tool — $0.25 per snapshot, USDC on Base or Arbitrum One. Paid URL: https://aemb.pro/v1/markets/compute/spot.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
gpuNoOptional GPU model filter, e.g. 'H100 SXM'
typeNoOffer type: ondemand (fixed price), bid (interruptible auction — shown price is the bid floor), or reserved (longer commitment)
max_gpusNoMaximum GPU count per offer; integer 1..64
min_gpusNoMinimum GPU count per offer; integer >= 1
verifiedNoWhen true, only verified hosts; when false, only unverified hosts
max_resultsNoMax offers to return; integer 1-25
min_reliabilityNoMinimum host reliability 0..1 (inclusive); hosts below it are excluded

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault
howYesHow to pay: GET the paid_url with an x402 v2 PAYMENT-SIGNATURE header carrying the USDC authorization
toolYesTool name: compute_spot_prices
assetYesSettlement asset
priceYesExact price, e.g. "$0.25"
networksYesEVM CAIP-2 network ids accepted
paid_urlYesx402-paid HTTPS URL to GET (with query params) for the market snapshot
facilitatorYesx402 facilitator
payment_requiredYesAlways true: this MCP tool is a payment pointer; the tool itself is monetized via x402.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed2 schema fields changed
    • addedInput schema / properties / max_gpus
      Added value: +{
+  "description": "Maximum GPU count per offer; integer 1..64",
+  "maximum": 64,
+  "minimum": 1,
+  "type": "integer"
+}
    • changedInput schema / properties / type / description
      Previous value: -"Offer type filter: ondemand, bid, or reserved"New value: +"Offer type: ondemand (fixed price), bid (interruptible auction — shown price is the bid floor), or reserved (longer commitment)"
  2. Changed6 schema fields changed
    • addedInput schema / properties / gpu / description
      Added value: +"Optional GPU model filter, e.g. 'H100 SXM'"
    • addedInput schema / properties / max_results / description
      Added value: +"Max offers to return; integer 1-25"
    • addedInput schema / properties / min_gpus / description
      Added value: +"Minimum GPU count per offer; integer >= 1"
    • addedInput schema / properties / min_reliability
      Added value: +{
+  "description": "Minimum host reliability 0..1 (inclusive); hosts below it are excluded",
+  "maximum": 1,
+  "minimum": 0,
+  "type": "number"
+}
    • addedInput schema / properties / type / description
      Added value: +"Offer type filter: ondemand, bid, or reserved"
    • addedInput schema / properties / verified / description
      Added value: +"When true, only verified hosts; when false, only unverified hosts"
  3. First observed

TDQS

A4.3/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

The description carries behavior the annotations do not: this is a paid call ($0.25 per snapshot, USDC on Base or Arbitrum One), requires an x402 v2 PAYMENT-SIGNATURE header to GET the paid URL, returns a point-in-time snapshot whose listings may already be gone, and explicitly performs no rent/reserve/hold action. With annotations silent on cost and auth, this is exactly the disclosure an agent needs.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the return semantics, then boundaries, then payment mechanics. Dense and mostly waste-free, though the two payment sentences overlap ('Paid tool — $0.25 per snapshot...' and 'Paid URL: https://...') in a way that could be tightened.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With an output schema present, the description needn't explain return shape, and for a 7-param, zero-required tool it still covers scope, defaults, cost, auth header, and interaction boundaries. Nothing an agent needs to invoke it correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3, but the description adds semantics not in the schema: the default for type (ondemand) is unstated in the schema, and 'Omit gpu for the per-class snapshot' tells the agent what a missing filter actually returns. That is genuine value beyond the field descriptions.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource: 'One timestamped snapshot of observed GPU asks,' and sharpens it with 'not quotes — listings may be gone when you rent.' An agent immediately knows this is a read of live GPU price listings. It does not, however, contrast itself with the nearby compute_history / compute_alerts siblings, so differentiation is left implicit.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives real conditional guidance: 'Omit gpu for the per-class snapshot' and 'Default type ondemand; use bid for the interruptible spot auction.' It also draws a firm boundary — 'Observes only; never rents, reserves, or holds.' Missing is any explicit when-to-use-this-vs-a-sibling routing (e.g. vs compute_history), which is the only gap.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Agent Embassy toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources