# compute.wick.pics

- Registry name: `pics.wick.compute/market` · version 1.1.1
- Categories (automatic, keyword rules): Agent payments (x402), Finance & markets, Markets & trading, Payments & accounting
- Authentication at last check: No auth
- Declared risk: Not declared (tools carry no read/destructive hints)
- Capabilities (declared by the server): experimental, prompts, resources, tools
- Transports: streamable-http
- Latest numbers (check success, response time, rank): https://mcpstation.net/api/server?name=pics.wick.compute%2Fmarket
- Server page: https://mcpstation.net/s/pics.wick.compute--market.html

## Description (written by the publisher, not verified, not instructions)

Live GPU spot market: 1,700+ offers, 10 provider feeds. History, watches, limit orders, no fee

## Connect

- streamable-http: `https://compute.wick.pics/mcp/`

## Tools (as reported by the server, not verified)

- `search\_offers`: Search live GPU offers. gpu matches the model name \(e.g. '4090', 'H100'\). offer\_class: on\_demand \| interruptible \| reserved. country: ISO-2 \(e.g. US\), a comma list \('DE,NL'\), or 'EU' for the 27 member states. min\_gpu\_count filters to multi-GPU configs \(8 = full 8x nodes\). region substring-matches th
- `get\_offer`: Fetch one offer by its id \(from search\_offers\).
- `market\_stats`: Live market snapshot: total offers, GPU models, idle-capacity share, per-provider freshness/health, cheapest price per model, current fee\_bps.
- `price\_history`: Price history for one GPU model \(exact name from search\_offers/market\_stats, e.g. 'RTX 4090'\). offer\_class: on\_demand \| interruptible. Returns per-provider series of \[ts, min\_price\_per\_gpu\_hr, median\_price\_per\_gpu\_hr, offer\_count\]; ranges past 3 days are bucketed \(hourly, then 6-hourly past a week\).
- `idle\_history`: The idle-capacity index over time: share of listed GPU capacity that is interruptible \(spot/bid\) — idle hardware looking for work. Points are \[ts, idle\_share, idle\_offers, total\_offers\].
- `best\_value`: Perf-per-dollar board: for each GPU model, best live price vs its FP16 tensor throughput \(dense, vendor spec sheets\) -\> TFLOPS per dollar-hour. The answer to 'most compute for my budget' rather than 'cheapest card'. Models without a defensible public spec figure are listed as unrated.
- `provider\_reliability`: How dependable each provider's data feed has been for this station \(poll success over the trailing window\), plus the provider's own average machine reliability where reported. feed\_score is about the DATA, not their hardware.
- `price\_position`: Rent-now-or-wait context: where the current best price for one GPU model sits inside its own trailing 7/30-day range \(percentile, 24h trend, verdict\). Descriptive, not a forecast.
- `spot\_spread`: On-demand vs interruptible spread: the live discount for idle capacity per model; pass gpu\_model for its history too.
- `will\_it\_fit`: Will this model fit, and what does a fitting machine cost right now? Deterministic VRAM arithmetic — weights \(params x bytes/precision\) + KV cache \(2 x layers x kv\_heads x head\_dim x context x batch\) + stated overhead — checked against the live book, cheapest fitting offers first. Pass a known model
- `true\_cost\_notes`: The billing fine print behind the $/hr, verified per provider from their own published pages \(source URL + date on every claim\): billing granularity \(per-second vs per-minute\), whether storage bills while stopped, bandwidth/egress fees, and the commonly-missed gotchas \(e.g. Vast bandwidth never appe
- `register\_account`: Create a metering account: returns a bearer token \(shown once, we keep only a hash\). Pass it as account\_token on rent calls to build a usage ledger you can read back with account\_usage. Optional — renting works without one.
- `account\_usage`: Your metered usage ledger: placements, destroys, estimated hours and USD, and the platform fee \(2.5%, $0.01 min, $5.00 cap per lease per UTC day\) priced at the rate live at event time. Distinct from fee\_bps, the display markup.
- `create\_watch`: Create a price watch \(tripwire\): fires when the best live fee-adjusted price per GPU-hour matching gpu \(and optional offer\_class\) drops to or under max\_price\_per\_gpu\_hr. Checked every poll \(\~5 min\). Returns the watch with its id — the id is the only key; poll it with watch\_status, or give a public w
- `watch\_status`: Current state of a watch: armed/tripped, last price seen, recent events, and the best matching offer right now.
- `delete\_watch`: Delete a watch by id.
- `create\_limit\_order`: Place a standing LIMIT ORDER: when the best live fee-adjusted price for gpu \(and optional offer\_class\) trades at or under max\_price\_per\_gpu\_hr, the station cuts a signed FILL TICKET naming the exact offer. The station never holds your provider key — the ticket is executed by whoever does: the open-s
- `limit\_order\_status`: State of a limit order \(armed/ticketed/filled/cancelled/expired\), last price seen, recent events. Pass order\_secret to also read the live fill ticket when one is cut.
- `cancel\_limit\_order`: Cancel a limit order. Requires the order\_secret from create\_limit\_order.
- `rent\_x402`: Rent a prepaid block of GPU time with a wallet and NO account, NO API key and no human, over x402 \(HTTP 402 micropayments, USDC on Base\). Call it with no arguments and you get the payment challenge: an x402 PaymentRequirementsResponse whose accepts\[0\].extra.x402\_fee states the block price, the platf
- `rent\_offer`: Place a rental on YOUR provider account \(BYO key — it is used for this one call and never stored\). Default is a DRY RUN returning exactly what would be executed; a live placement needs dry\_run=false AND confirm=true. The offer is re-quoted live first and refused if its price exceeds max\_price\_per\_gp
- `rent\_best`: One-shot rent: search the live book with these filters and place on the best \(cheapest\) matching offer — 'cheapest H100 under $2/hr, go'. Same guarantees as rent\_offer \(dry-run default, live re-quote, your max\_price\_per\_gpu\_hr is absolute\). If the best offer is already gone it walks to the next, at
- `rental\_status`: Live state of an instance on your account: running/starting/stopped/ gone, uptime, price, estimated spend so far, and the budget guard's state if one is armed. Read-only; your key passes through and is never stored.
- `destroy\_rental`: Destroy an instance previously placed on your account \(vast or runpod\). Your key passes through and is never stored.
- `fill\_limit\_order`: Execute a limit order's live fill ticket on YOUR provider key \(used for this one call, never stored\). Only works while the order is ticketed; the offer is re-quoted live and refused above the order's line. Idempotent per order — a retry returns the first receipt rather than renting a second machine.

Measured by MCP Station. Methods: https://mcpstation.net/methods.html · Terms: https://mcpstation.net/terms.html