GPU Finder MCP server
OfficialGPU Finder MCP server lets an agent query cloud GPU models, providers, live prices, stock availability, and price history through read-only tools.
List tracked GPU models with specs, provider/instance counts, and hourly price ranges.
List cloud GPU providers with GPU counts, price ranges, egress costs, links, and update times.
Get every provider offer for a GPU and node size, including on-demand and spot USD per whole-node hour.
Find the cheapest listed offer and cheapest recently in-stock offer via market snapshot.
Check observed stock availability by provider and node size over the last seven days.
Retrieve up to 12 months of monthly price history per provider.
Check data freshness and per-provider collection status.
Check API key usage and monthly allowance without spending units.
Use a rental prompt and pricing/stock guide to interpret whole-node and per-GPU prices.
All tools are read-only; data calls use monthly allowance units, while get_usage spends none.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@GPU Finder MCP server What's the cheapest 8x H100 node in stock right now?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
GPU Finder MCP server
Live cloud GPU prices, stock and price history across 30+ providers (hyperscalers, neoclouds and marketplaces), for Claude, Cursor, VS Code, Cline and any MCP client.
Ask your agent things like:
"What is the cheapest 8× H100 node in stock right now, and on which provider?"
"Where can I rent a single RTX 5090 today, and for how much per hour?"
"How has B200 pricing moved over the last six months?"
The data comes from GPU Finder, which re-checks stock every hour.
Get a key
Create a free API key (60 data calls a month) at https://gpufinder.dev/developers. Paid plans add more calls.
Related MCP server: @inferra-trade/mcp-server
Install
npx -y gpufinder mcp runs the server over stdio. Add it to your client:
Claude Desktop / Cursor / Cline / Windsurf (claude_desktop_config.json, .cursor/mcp.json, …):
{
"mcpServers": {
"gpufinder": {
"command": "npx",
"args": ["-y", "gpufinder", "mcp"],
"env": { "GPUFINDER_API_KEY": "YOUR_API_KEY" }
}
}
}Claude Code:
claude mcp add gpufinder --env GPUFINDER_API_KEY=YOUR_API_KEY -- npx -y gpufinder mcpVS Code (.vscode/mcp.json):
{
"inputs": [
{ "type": "promptString", "id": "gpufinder-api-key", "description": "GPU Finder API key", "password": true }
],
"servers": {
"gpufinder": {
"type": "stdio",
"command": "npx",
"args": ["-y", "gpufinder", "mcp"],
"env": { "GPUFINDER_API_KEY": "${input:gpufinder-api-key}" }
}
}
}Requires Node.js 20 or newer. Keep your key out of files you commit.
Prefer no install?
Clients that support remote MCP servers can connect straight to https://gpufinder.dev/mcp with the header
Authorization: Bearer YOUR_API_KEY. See https://gpufinder.dev/api.
Tools
All tools are read-only.
Tool | What it returns |
| GPU models tracked, with specs and price range |
| Providers with GPU counts and links |
| Per-provider offers for a GPU and node size (USD per hour) |
| Cheapest listed offer and cheapest offer observed in stock at the last check |
| Observed stock by provider |
| Up to 12 months of price history |
| Data freshness per provider |
| Your monthly allowance and usage (free) |
Listing tools is free. Each data call uses one unit of your monthly allowance; get_usage uses none.
Rental prompt and pricing guide
Choose find_gpu_rental in your MCP client and enter a GPU model and the number of GPUs per node. For example, enter H100 and 8 to compare the cheapest listed eight-GPU node with the cheapest one recently observed in stock.
Read gpufinder://guides/pricing-and-stock for an example of whole-node and per-GPU hourly prices and how to interpret stock observations. Opening the prompt or guide is free. Running the market snapshot uses one data unit.
Environment variables
Variable | Purpose |
| Your API key (required for tool calls) |
| Override the endpoint (default |
| Log requests to stderr; the key is never logged |
How it works
This package is a small stdio proxy with no dependencies. It forwards each MCP message to the hosted GPU Finder endpoint with your key and relays the reply, so the tool list always matches the live service.
Also listed in the official MCP Registry as dev.gpufinder/gpufinder.
License
MIT
Available Tools
8 toolsget_availabilityARead-onlyInspect
Show day-by-day stock for one GPU model over the last seven days, per provider and node size: instances observed available versus total each day. Optionally limit to one provider. Use it to judge how reliably a GPU can be found in stock; prices come from get_prices. Days without observations are unknown, not zero. Pages default to 20 rows, maximum 100; each page costs one data unit.
Inputs: gpu (required), page, provider.
get_data_statusARead-onlyInspect
Report how fresh GPU Finder's data is: when price and availability collection last completed, and each provider's last success, last run, consecutive failures and last error. Use it to decide whether a price or availability answer may be stale. Costs one data unit; it is not an anonymous health check.
Inputs: no named parameters.
get_market_snapshotARead-onlyInspect
Answer "what is the cheapest way to rent this GPU now": the cheapest listed offer and the cheapest offer observed in stock for one GPU model and node size, each with provider, region, whole node and per GPU hourly USD price, availability status and check time, plus how many providers were compared. Use it for a single recommendation; use get_prices for every offer. The whole snapshot costs one data unit; collected stock is not a capacity guarantee.
Inputs: gpu (required), count.
get_price_historyARead-onlyInspect
Show how one GPU model's price changed month by month per provider: on-demand and spot USD per whole node hour for up to 12 UTC months. Use it for trends; use get_prices for current prices. Pages default to 20 rows, maximum 100; each page costs one data unit. Reuse the same filters and page limit with nextCursor.
Inputs: gpu (required), page, count, months.
get_pricesARead-onlyInspect
List every provider offer for one GPU model and node size: provider, instance type, region, GPU count, on-demand and spot USD price per whole node hour, links and last update. Use it for the full price table; use get_market_snapshot for only the cheapest offer, get_availability for stock and get_price_history for trends. Prices reflect collected data, not a live capacity guarantee. Pages default to 20 rows, maximum 100; each page costs one data unit.
Inputs: gpu (required), page, count.
get_usageARead-onlyInspect
Show this API key's usage across REST and MCP: plan allowance, units used and remaining, and UTC reset times. Use it before large paginated reads or after a quota_exceeded error. Spends no monthly units and works after the monthly allowance is exhausted, but is still rate limited.
Inputs: no named parameters.
list_gpusARead-onlyInspect
List the GPU models GPU Finder tracks: canonical name and slug, VRAM, memory type, bandwidth, interconnect, provider and instance counts, and min/max hourly USD price. Use it to find valid gpu names for the other tools or to compare GPU specs; use get_prices for per-provider offers. Pages default to 20 rows, maximum 100; each page costs one data unit.
Inputs: page.
list_providersARead-onlyInspect
List GPU cloud providers: name and slug, GPU and instance counts, min/max hourly USD price, egress cost per GB and free egress allowance, website and GPU Finder links, and last update time. Use it to find valid provider slugs or compare providers overall; use get_prices for one GPU's offers. Pages default to 20 rows, maximum 100; each page costs one data unit.
Inputs: page.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 8 tools
Each tool targets a clearly distinct information need: catalog lists, full price tables, stock history, price trends, cheapest recommendation, data freshness, and API usage. The descriptions explicitly cross-reference neighboring tools (e.g. get_prices vs get_market_snapshot vs get_availability) to prevent misselection.
All tool names follow a consistent snake_case verb_noun pattern using get_ or list_. The nouns are specific and predictable across the set.
Eight tools is well-scoped for a read-only GPU pricing and availability service. There are no redundant tools, and each operation earns its place.
The surface covers discovery (GPUs/providers), current prices, availability, historical trends, cheapest-offer recommendation, data freshness, and usage/quota. Minor gaps like a dedicated single-GPU lookup or provider-specific GPU listing are workable through existing list and price tools.
Maintenance
Related MCP Connectors
Live GPU rental and LLM inference prices across dozens of providers, plus GPU price forecasts.
Compare live GPU cloud rental prices and match workloads to the cheapest provider.
Live public cloud GPU rental prices (H100, A100, L40S and more) across 20+ providers.
Cloud GPU prices with live stock from 22 providers. Find where to rent an H100 or B300 now.
Related MCP Servers
- AlicenseAqualityBmaintenanceGlobal price benchmarking for AI inference across 2,600+ SKUs from 47 vendors. Query live pricing, market indexes, and model specs via 8 tools. Free tier available.859 npmMIT
- AlicenseAqualityCmaintenanceEnables AI agents to retrieve live GPU rental prices, compare models, and access market summaries and historical price data through four read-only MCP tools.4MIT

gpu.nz MCP serverofficial
AlicenseNot gradedqualityCmaintenanceEnables AI agents to search live GPU offers, estimate model VRAM, rent and monitor cloud GPU servers with ready templates, read logs, and stop or destroy them when done. Browsing and VRAM estimates work without an API key, while server management requires one.1MIT- AlicenseNot gradedqualityCmaintenanceEnables agents to query real-time GPU rental prices across 26 platforms, including current lowest quotes, 30-day price history, market movers, platform reliability, and task/budget recommendations. Access is read-only and metered per call via x402 with USDC on Base, requiring no signup or API key.MIT