@jam-jam/convert-mcp

Separate a song into stems

jj_stems_separate

Upload a local audio file to start asynchronous stem separation into vocals and instrumental, or four stems (drums, bass, other); then poll the job and download the results.

Instructions

Uploads a local audio file (path, or base64 content) and starts its separation into vocals + instrumental (stems=2) or vocals + drums + bass + other (stems=4) (POST /stem-tool/jobs). Asynchronous and slow (minutes, up to hours for a long file): by default this returns at once with the job, then follow it with jj_stems_get_job and fetch the files with jj_stems_get_file; pass maxWaitMs to wait for a short file instead. Needs the stems:write scope and an API key from a Premium account (403 otherwise). Limits on Premium: a file of at most 1 h 30 and under 2 GB (5 minutes and 100 MB without Premium). Quotas, in UTC days and calendar months, with ONE counter per account shared by both models and by the website: 30 jobs/day and 500/month on Premium (81000 s of audio/day, 540000 s/month). Only one separation can run at a time per account. A 429 states when the counter resets; a 413 means the file is too large or too long. Stems come as MP3 320 kb/s by default, or as FLAC 24-bit with format="flac" (lossless, about 4 times larger). Check jj_stems_capabilities first.

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": "jj_stems_separate"
  }
}

Input Schema

Name	Required	Description
`path`	No	Absolute or relative local path to the audio file. Mutually exclusive with contentBase64.
`model`	No	"fast" (default) or "quality" (SCNet-large, slower, same quota).
`stems`	No	2 = vocals + instrumental (default); 4 = vocals + drums + bass + other.
`format`	No	"mp3" (default, 320 kb/s) or "flac" (24-bit lossless, for mixing or mastering).
`filename`	No	File name (with extension). Required with contentBase64; defaults to the basename of "path".
`maxWaitMs`	No	Wait up to this long (milliseconds, cap 600000) for the separation to finish. Default 0: return at once. Waiting needs a key with convert:read, otherwise the job is returned at once. The job keeps running if it is exceeded.
`contentBase64`	No	Base64-encoded audio content. Mutually exclusive with path. Requires "filename".

Schema Changelog

Changes observed during successful MCP inspections.

First observedv0.3.2Nov 13, 2026

TDQS

A4.8/5.0

Behavior5/5

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

With no annotations the description carries the full burden and does so: asynchronous and slow (minutes to hours), one separation per account at a time, exact quota counters (30/day, 500/month, shared across models and website), Premium vs non-Premium size/duration limits, and the meaning of 403, 429 and 413 responses.

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?

Dense but well-organized, front-loading what the tool does and then behavior, limits, formats and errors. The one structural weakness is that the prerequisite 'Check jj_stems_capabilities first' sits at the very end instead of leading, and the single paragraph packs many facts together.

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?

For an async, quota-limited, auth-gated job-creation tool with no output schema, it covers everything an agent needs: return shape (a job), follow-up tools, auth scope, account tier, quotas, concurrency, file limits, output formats and error semantics.

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?

Schema coverage is 100% so the baseline is 3, but the description adds genuine meaning beyond the schema: FLAC is lossless and about 4x larger, MP3 is 320 kb/s, model choice shares the same quota, and maxWaitMs is framed as 'wait for a short file' rather than the default return-at-once behavior.

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?

Names a specific verb and resource ('separates a local audio file into stems'), states the endpoint (POST /stem-tool/jobs), and disambiguates the two modes (stems=2 vocals+instrumental vs stems=4 vocals+drums+bass+other). It also makes clear this is the job-creation step, distinct from jj_stems_get_job and jj_stems_get_file.

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?

Explicitly routes the agent: default returns immediately and you follow with jj_stems_get_job / jj_stems_get_file, or pass maxWaitMs for a short file instead. It also states the prerequisite 'Check jj_stems_capabilities first' and the Premium/scope requirement, so when-to-use and dependencies are covered.

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