Skip to main content
Glama

Upload a CPET recording

upload_cpet

Upload a CPET export whose contents are already in the conversation (pasted or attached) and get a cpet_id back.

The file is sent exactly as exported, unchanged: Oxynet does the parsing. It detects the vendor format and knows its unit conventions, per-phase clock resets and decimal-separator defects. A file parsed or converted beforehand loses that (a VO2 column in L/min read as mL/min is off by 1000, and nothing downstream notices).

For a file on disk, a call with no arguments returns the same one-time upload link as create_upload, so the bytes go from disk to Oxynet without entering the conversation (a 360 KB export is about 97,000 tokens as text).

UNSUPPORTED_FORMAT means no parser recognised the file. Its recovery block gives the canonical JSON shape, the required channels and units, and the formats considered, for a file converted into that shape.

Returns cpet_id (the handle every other tool takes), the detected source format, sampling as the file arrived, channels with their coverage, and flags. The recording stays on the server.

Privacy: the uploaded bytes are parsed and discarded, never stored, and the parsed record is deleted after 24 h. Oxynet needs no patient identifiers; CPET exports often carry the patient's name in the filename and header.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "upload_cpet"
  }
}

Input Schema

NameRequiredDescriptionDefault
fileNo
contentNoThe file's contents, verbatim: text for CSV/TXT/XML/JSON, base64 of the bytes for XLS/XLSX.
filenameNoAny name ending in the file's real extension, which selects the parser. Only the extension is used; the name is not stored.
retain_hoursNoHours to keep the parsed record beyond the 24 h default.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed4 schema fields changed
    • changedInput schema / $defs / AttachedFile / description
      Previous value: -"A file attached in ChatGPT, passed by reference."New value: +"A file attached in ChatGPT, passed by reference and downloaded by Oxynet, so the bytes never pass through the conversation. Takes the place of `content`."
    • addedInput schema / properties / content / description
      Added value: +"The file's contents, verbatim: text for CSV/TXT/XML/JSON, base64 of the bytes for XLS/XLSX."
    • addedInput schema / properties / filename / description
      Added value: +"Any name ending in the file's real extension, which selects the parser. Only the extension is used; the name is not stored."
    • addedInput schema / properties / retain_hours / description
      Added value: +"Hours to keep the parsed record beyond the 24 h default."
  2. First observed

TDQS

A4.4/5.0
Behavior3/5

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

Annotations already declare the write/not-idempotent/closed-world profile, and the description adds substantial context beyond that: bytes are sent verbatim, Oxynet does the parsing, the UNSUPPORTED_FORMAT error carries a `recovery` block, and uploaded bytes are discarded with the parsed record deleted after 24 h. It stops short of describing idempotency or deduplication behavior (re-uploading the same file) and does not mention the `retain_hours` override that the schema exposes against the stated 24 h default.

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-loads the action, the invariant (file unchanged), and the return value before edge cases and privacy. Dense but nearly every sentence carries decision-relevant information; the token-cost aside (360 KB ≈ 97,000 tokens) is illustrative rather than wasteful, though the block is longer than strictly needed.

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?

No output schema exists, and the description covers the return payload (`cpet_id`, detected `source`, `sampling`, `channels` with coverage, `flags`) plus the failure mode and its recovery payload. Combined with the privacy and retention notes, an agent has everything needed to call it and handle the error path.

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?

With 75% schema coverage the baseline is a 3; the description earns more by clarifying the mutually exclusive `content`/`file` paths and that the zero-argument form yields an upload link instead of a parse. It adds nothing about `retain_hours`, leaving the 24 h statement in tension with that parameter for an agent skimming only the prose.

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

Purpose5/5

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

States a specific verb and resource (upload a CPET export), the precondition (contents already in the conversation), and the return handle (`cpet_id`). It explicitly distinguishes the on-disk case from the sibling `create_upload`, so an agent can separate the two without opening either schema.

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

Usage Guidelines5/5

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

Gives explicit when-to-use (file already pasted/attached) and when-not guidance, naming the alternative path for files on disk (zero-argument call returning the same one-time upload link as create_upload). The VO2 unit-mismatch example makes concrete why pre-converted input is disallowed.

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

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources