Find a tool for two overlapping voices.

What you receive

Expected result in this snapshot: one SpeechBrain record. The complete public example is below. Its source-review date is 2026-10-16. Catalogue source review: 2026-10-19. Neither date is today’s execution date or a promise of daily updates.

{
  "id": "github:speechbrain/speechbrain",
  "name": "SpeechBrain: separate two overlapping voices",
  "purpose": "Separate a two-speaker audio mixture into two replayable WAV tracks.",
  "category": "Audio & recordings",
  "type": "Python speech separation component",
  "platforms": [
    "Python setup component",
    "Local CPU; GPU optional",
    "Separate model download required"
  ],
  "sourceURL": "https://github.com/speechbrain/speechbrain",
  "getSoftwareURL": "https://github.com/speechbrain/speechbrain/blob/c0241278f93ef02cd5191f03caf4e5f8e1292bf6/README.md",
  "licenseDeclaration": "Apache-2.0",
  "licenseURL": "https://github.com/speechbrain/speechbrain/blob/c0241278f93ef02cd5191f03caf4e5f8e1292bf6/LICENSE",
  "requirements": "Setup component, not a finished recorder app. Reviewed SpeechBrain 1.1.2 release/tag c0241278f93ef02cd5191f03caf4e5f8e1292bf6 equals current source (2026-10-14). Python >=3.8.1, matching PyTorch/torchaudio >=2.7.1, soundfile >=0.12.1, HyperPyYAML and other declared dependencies. PyPI publishes the package and matching torch/torchaudio 2.7.1 wheels for Python 3.10-3.13 on macOS arm64. CPU is supported in the actual interface; CUDA is optional. No automatic Apple MPS or speed claim. The ungated Apache-2.0-declared speechbrain/sepformer-wsj02mix checkpoint is a separate first-use download (~113 MB repository storage), not fetched in this review. Narrow route: mono, two-speaker mixture at 8 kHz. It was trained on WSJ0-2Mix; quality on meetings, noisy recordings, accents or more speakers is not established. Output order need not identify the human speakers. Primary model card saves using torchaudio; current SDK also implements soundfile-based audio_io.save. No audio, dependencies or checkpoints installed/executed; source verification only.",
  "evidenceLimits": [
    "Give it a mono recording containing two overlapping speakers and receive two separate 8 kHz WAV files. This supplies separate audio tracks to replay or pass to an existing transcription tool, removing manual attempts to untangle both voices in one mixed track. It separates waveforms; it does not name speakers, transcribe them or establish evidentiary accuracy.",
    "Source review only; software was not installed or run, and individual accessibility fit was not tested."
  ],
  "reviewedAt": "2026-10-16"
}

This is a static result computed locally from the current catalogue, not a new paid delivery. Future catalogue changes can change results.

Website: use your direct subscription

Direct access costs $5.99 per month for 1,000 admitted searches. After purchase, use the direct subscription key. A marketplace key will not work here.

curl --get "https://www.myworkingmemory.ai/api/software-library/direct/search" --header "Authorization: Bearer $CURATED_FINDS_API_KEY" --data-urlencode "query=SpeechBrain" --data-urlencode "limit=1" --data-urlencode "offset=0"

Read the returned records array. The direct response also includes allowance and pagination fields. This example requests no second page.

Apify: pay for one completed page

Open Curated Finds on Apify using your paid Apify account. Paste this into Input. Check the price and set a finite maximum charge before choosing Start. Nothing here starts a run.

{
  "query": "SpeechBrain",
  "limit": 1,
  "offset": 0
}

One completed catalogue page costs $0.00599, even if a valid search is empty. Up to 25 records may fit on a page; this example requests one. This is not a monthly 1,000-call subscription. Separate platform operating/storage charges and your plan can apply; a run’s event-charge limit does not cap every operating expense.

After a successful run, open its Dataset and view or export the record. OUTPUT also contains the version and page metadata. For Make, use the existing guide above: connect both modules to your own account, keep On demand, and receive the records in module 2. Sheets, email and Slack delivery are not included.

MCPize: search inside your connected client

Choose the $5.99/month, 1,000-request Pro plan on MCPize, then use its connection instructions in your client. New Free enrollment is inactive.

Call search_software with exactly these arguments:

{
  "query": "SpeechBrain",
  "limit": 1,
  "offset": 0
}

Inspect the tool result’s records. The other native tools are get_software and catalogue_info. These names are specific to MCPize; they are not Apify tool names. Native request accounting is not a count of useful delivered records.

RapidAPI: use its generated request

Open the RapidAPI offer. Pro is $5.99/month for 1,000 requests; Basic allows one request per month. The separate bandwidth allowance and overage terms still apply.

Select the search endpoint and enter these query parameters. Keep the gateway host and authentication headers from RapidAPI’s generated code sample, using your own selected application.

GET /search?query=SpeechBrain&limit=1&offset=0

Inspect records in the response. Do not paste the direct subscription key into this gateway.

API.market: use its own key and search endpoint

Open the API.market offer. Pro is $5.99/month for 1,000 hard-capped calls. Its 50-call, seven-day Free trial is separate.

Select search and use the platform’s generated host and authentication example with these query parameters:

GET /search?query=SpeechBrain&limit=1&offset=0

Inspect records. If you later retrieve a record by ID here, use /records?id=…, not a slash-containing record ID in the URL path. That would be another request, outside this one-search example.

How to tell whether it worked

The returned record ID must be github:speechbrain/speechbrain, with the same source-review date and 13 public fields as above. Search responses and Apify Dataset exports have different outer wrappers; compare the record itself. Keep the version/run reference privately if you need support.

Do not automatically retry a timeout or empty result. Check the existing run, output and allowance first; an earlier attempt may already be billed. One platform purchase does not unlock the others.

Owner tests establish technical delivery only. Independent customer purchases, actual seller proceeds, refunds and fully allocated costs remain separate evidence. Sustainable profit is not yet established.

Snapshot version: 6c4567c485c357318c2722eaaaa0d03aae6d6249c663080346bf601d24cae4a0. No secrets, paid calls, checkout, schedule or analytics events are created by this page.