AgentCompute

Read a previously submitted job by job_id. Returns status, transcript, runtime and reference quality when complete. For queued or running jobs, wait poll_after_seconds before checking again. Failed or expired jobs are not successful results.

Check whether the GPU provider can accept audio transcription work. Returns availability, permitted task types, and limits. Does not start a job.

Public live availability, model catalog, free-trial limits and polling guidance. No token required. Check before submitting; an offline GPU cannot execute work.

Return the completed image as an MCP image content block. Requires the same token/identity that submitted the job. Images expire after 24 hours from submission; download before expiry.

Read your own chat/image job. Includes partial text while running, final text or image availability, GPU runtime and server timings. Wait poll_after_seconds between checks; stop on succeeded, failed or expired. Does not create or retry work.

Experimental local worker agent. Version 1 reads supplied documents. Set agent_task.version=2 for decimal calculations, CSV preview/filter/numeric sort/summary/group sums, and TXT/Markdown/CSV/JSON files using approved local CPU/RAM functions. Check services.agent_tools for version 2. Up to eight tool calls, eight optional documents, 24000 input characters and three files (16000 bytes each, 24000 bytes total). Set required_artifacts to exact filenames to require delivery. Version 3 requires services.agent_web and adds public HTTPS page reading and web search, with source URLs. No JavaScript browsing, logins, VM, arbitrary code or shell. Search availability and website access can vary. File contents are returned in result.agent_result.artifacts. Quotes are matched; tool outputs and factual correctness are not independently verified. Shared trial quotas. Poll get_studio_job.

Experimental structured extraction from one to six nonconfidential product records supplied as text. No URL fetching or OCR. Check catalog_extraction.available first. One task unit, bounded retries; retain idempotency_key. Output includes JSON/CSV with source quotes and null for missing data. Schema and source matching are checked, semantic correctness is not guaranteed. Poll get_studio_job.

Default routing fairly rotates compatible idle providers. Optional parallel:{parts:2|3,objective,documents} splits document analysis and returns one synthesized result under the same job ID. Each part and synthesis count toward quotas. Otherwise provide messages. Routing fastest requires comparable configured benchmarks; lowest_cost uses configured prices, with fair fallback when metrics are unknown. No payment. Queue a free text task: draft, summarize, extract, translate text, reason or assist with code. No code is executed. Provide user/assistant messages, explicit consent for nonconfidential permitted content, and a stable idempotency key. Returns immediately with a job ID. Poll get_studio_job. Uses the configured local model, no external AI fallback.

Beta users must supply their uploaded asset_id and kind compute_job. No accuracy guarantee is made for uploaded files. Owner legacy calls can queue the fixed synthetic benchmark. compute_job returns a transcript; delegated_task additionally requires word error rate at most 10% against the known reference. Returns a job_id to poll with get_compute_job. No payment is charged. Reuse the same idempotency_key for retries, and use a new key only for a new job.

Queue a free text-to-image task from a complete scene description. Each task is independent, not an edit of a previous image. Returns a job ID; poll get_studio_job then get_studio_image for the PNG. Configured model and size appear in service status/results. No external AI fallback.