canonical: https://jentic.com/apis/liquidcompute.com/liquidcompute # LiquidCompute API The LiquidCompute API serves read-only GPU rental price data in USD per GPU-hour: custom, transaction, and informational price indexes for GPU families such as Hopper and Blackwell, filterable by region group and contract type. It also returns the latest vendor offers from its price scrapes, daily price history and change logs for models from the A100 to the B300, individual private transaction records, and commitment pricing by contract length. Reference series with history and pairwise correlation, plus a dataset freshness report, help put GPU prices in context. ## For AI agents Read GPU price indexes, current and historical vendor GPU offers, private transaction records, and commitment pricing, all in USD per GPU-hour. ## Scope Does not rent, provision, or manage GPU instances, and does not place trades. Use for reading GPU price indexes, vendor offers, and transaction data only. ## Capabilities - Get the latest custom GPU index value for a region group, contract type, and vendor category - Read daily transaction index values for Hopper or Blackwell GPUs by region and spot, on-demand, or combined contracts - Fetch the current vendor GPU offers from the newest completed price scrape - Trace one GPU model's daily vendor prices and every price move, new listing, and withdrawal since a date - List private GPU transaction records by model, contract type, and region - Get commitment pricing by contract length for reserved GPU deals - Correlate two reference series and check when each dataset last updated ## Use cases ### Agent-assisted GPU procurement through Jentic An AI agent helping a team rent GPUs can compare current vendor offers for a model against the market index before recommending a provider. Through Jentic it finds the GPU price and index operations by intent and loads their schemas, so the developer does not hand-wire the X-API-Key header or the enum-heavy query parameters. Example prompt: Fetch the current GPU offers, filter them to H100, and compare the lowest USD per GPU-hour rate with the current transaction index for gpu_family hopper, region_group US, and contract_type combined. ### GPU price trend reporting Finance and infrastructure teams tracking compute costs can pull daily transaction index history, informational index history for an index_name such as hopper or blackwell, and custom index history at 30m, 1h, or 1d resolution. Intraday resolutions require the intraday entitlement. Example prompt: Get the informational index history for index_name blackwell in region_group US from start_date 2026-01-01 and summarize the change. ### Vendor price change alerts The change log for a GPU model lists every price move, new listing, and withdrawal revealed by a completed price scrape started on or after a given date, newest first, so an agent can flag when a provider cuts prices on A100, H100, H200, B200, or B300 offers. Example prompt: Get the price changes for gpu_model b200 since 2026-10-01 and list the vendors that lowered their price. ### Reserved contract benchmarking Before signing a reserved GPU contract, compare the quote with recent negotiated deals: commitment pricing returns recent reserved, non-rolling deals by contract length, and the private transactions list returns individual records by model, contract type, and region. Both depend on your organization's entitlements. Example prompt: List private transactions for gpu_model h100 with contract_type reserved and region us, then fetch commitment pricing and compare the two. ## Key endpoints | Method | Path | Description | | --- | --- | --- | | GET | `/v1/custom-index/current` | Get latest custom index value | | GET | `/v1/txn-index/current` | Get latest transaction index value | | GET | `/v1/gpu-prices/current` | Get current GPU prices | | GET | `/v1/gpu-prices/changes` | Get GPU price changes | | GET | `/v1/private-transactions` | List private transactions | | GET | `/v1/custom-index/commitment-pricing` | Get commitment pricing by contract length | | GET | `/v1/info-index/history` | Get informational index history | ## Key resources - **Custom index** — Custom index series, current and as-of values, history, commitment pricing, and reference - **Transaction index** — Daily transaction index values by GPU family, region, and contract type - **Informational index** — Hopper and Blackwell informational indexes with history - **GPU prices** — Current vendor offers, daily price history, and change logs per GPU model - **Private transactions** — Individual private GPU transaction records - **Reference series** — Reference series with history and correlation, plus dataset freshness ## AI readiness This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use. - **Score:** 66 / 100 - **Maturity:** AI-Aware - **Dimensions:** - Foundational Compliance: 99 / 100 - Developer Experience & Jentic Compatibility: 74 / 100 - AI-Readiness & Agent Experience: 47 / 100 - Agent Usability: 94 / 100 - Security: 50 / 100 - AI Discoverability: 69 / 100 - **View full report:** https://jentic.com/apis/liquidcompute.com/liquidcompute/scorecard - **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/ - **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview ### Score it yourself Every API in the directory is allowlisted, so you can re-score it with no key required. - **Score your own API:** https://jentic.com/scorecard.md - **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md ```sh npx @jentic/api-scorecard-cli score ``` ## Why Jentic - **Setup:** Wiring LiquidCompute by hand means sending the X-API-Key header, matching each index's enum filters such as region_group and contract_type, and honoring token-bucket Retry-After headers. Through Jentic you install once, import LiquidCompute from the API Directory, store the key once, and your agent calls it. - **Permission scoping:** LiquidCompute selects data with query parameters rather than path ids, so rules bound which operations your agent may call, not which GPU model or region. You choose the operations it may call, so entitlement-gated datasets like private transactions are not included unless you add them. - **Credential handling:** Your LiquidCompute API key is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'current H100 rental price' or 'GPU price changes', and Jentic returns the matching LiquidCompute operation with its input schema so the agent calls the right endpoint without browsing the reference docs. ## Related APIs - **Lambda Cloud API** — Launch and manage GPU cloud instances - **Paperspace API** — GPU machines and notebooks in the cloud - **Alpha Vantage API** — Stock, forex, and commodity market data - **EIA Open Data API** — US energy and electricity price data ## FAQ ### What authentication does the LiquidCompute API use? Per its OpenAPI spec the LiquidCompute API takes an API key in the X-API-Key header on its /v1 operations; missing or invalid keys return 401. The health check is public and needs no key. Which index families your key can read depends on your organization's entitlements. Through Jentic the key is stored encrypted by your own Jentic One instance and injected at request time. ### Is there a LiquidCompute MCP server? You don't need an MCP server to give your agent the LiquidCompute API. Jentic connects it directly from the API Directory: import it, store your API key once, and your agent discovers the index and GPU price operations on demand. ### Can I get current H100 prices with the LiquidCompute API? Yes. The current GPU prices operation returns the comparable vendor offers from the newest completed price scrape, normalized to USD per GPU-hour. For one model over time, the GPU price history operation takes gpu_model, such as h100, with a start and end date, and the changes operation lists price moves since a date. ### What are the rate limits for the LiquidCompute API? Per the spec, limits are token buckets per API key: you can burst up to your bucket size and sustain your refill rate. The X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers report your bucket, and Retry-After is set on 429 responses. The spec does not publish the bucket sizes; see https://docs.liquidcompute.com for your plan. ### Can I limit what my agent is allowed to do with the LiquidCompute API? Yes. The API is read-only, so the question is which data your agent may read. For example, allow only the current GPU prices and transaction index operations, so the agent can answer pricing questions while private transactions and commitment pricing stay outside its reach. Every call it makes is logged. ### How do I read the GPU transaction index through Jentic? Search Jentic for 'get the GPU transaction index', load the current transaction index operation, and pass filters such as gpu_family hopper, region_group US, and contract_type combined. Jentic injects your X-API-Key at call time. To run it on your own infrastructure, install Jentic One from its GitHub repo.