Public registry snapshot: 2026-10-03. Enable JavaScript or follow Public model JSON for current registry data.
PIPELINE SETUP
Voice sample quality and similarity enrichment
Attach speech-quality features and speaker/audio embeddings to catalog samples for comparison and similarity discovery.
Internal integration · public test unavailable1. Prepare inputs
The internal adapter takes existing Voice Models catalog sample identities, rather than a WAV upload.
- Sample identity
weights_id,base_sample_gender,pitch- Selection
- Active, available catalog samples; at most 18 per internal batch. Core resolves audio, hashes and decoded durations.
- Your own audio
- No public upload or sample-admission contract for this pipeline.
Component contracts (4)
- NISQA speech qualityquality features
Audio decoded and preprocessed according to upstream model documentation.
Upstream input documentation ↗ - Gender classifieroptional gender observation
Input limits and a public execution adapter are not registered.
- SpeechBrain ECAPAspeaker embedding
Input limits and a public execution adapter are not registered.
- LAION larger CLAP generalaudio embedding
Input limits and a public execution adapter are not registered.
2. Check the live contract
Quality features, optional gender observations and speaker/audio embeddings.
No public pipeline recipe is connected. This executable check resolves component revisions and declared licences; it does not run the models.
API quickstart · no key required
Bash, curl and jq. Run locally; only public GET requests are sent.
set -euo pipefail
CORE="${MODELUSE_CORE:-https://core.modeluse.com}"
PIPELINE_ID='voice-model-sample-enrichment'
CHECK_DIR=$(mktemp -d)
trap 'rm -rf "$CHECK_DIR"' EXIT
curl --fail --silent --show-error --max-time 20 "$CORE/v1/pipelines" > "$CHECK_DIR/pipelines.json"
curl --fail --silent --show-error --max-time 20 "$CORE/v1/models" > "$CHECK_DIR/models.json"
jq -en --arg id "$PIPELINE_ID" \
--slurpfile pipelines "$CHECK_DIR/pipelines.json" \
--slurpfile models "$CHECK_DIR/models.json" '
($pipelines[0].items | map(select(.id==$id)) | first) as $p |
if $p == null then error("Pipeline not in live registry") else
{pipeline_id:$p.id, availability:$p.availability,
public_execution_supported_by_this_client:false,
input:$p.input, expected_output:$p.output,
components:[$p.components[] | . as $c |
($models[0].items | map(select(.id==$c.model_id)) | first) as $m |
{component:($c.model_id // $c.inventory_name), role:$c.role,
source:$m.source_url, revision:$m.versions[0].revision,
declared_license:$m.versions[0].license,
input_contract:$m.versions[0].source_details.usage_contract.input}]} end'
# This checks real contracts. This pipeline has no public executable recipe.
Inspect published sample evidence ↗Existing Voice Models output; this does not start a new test.
Existing integration status
Read only, for your current workspace. Sign-in does not enable pipeline submission.
Read status through the API
Bash, curl, jq and your existing workspace API key. This GET neither admits work nor grants access.
set -euo pipefail
: "${MODELUSE_TOKEN:?Use your existing workspace API key}"
CORE="${MODELUSE_CORE:-https://core.modeluse.com}"
curl --fail --silent --show-error --max-time 20 \
-H "Authorization: Bearer $MODELUSE_TOKEN" "$CORE/v1/enrichment/status" \
| jq -e 'select(.schema=="modeluse-enrichment-status-v1" and .scope=="authenticated tenant") |
{scope,observed_at,selector_snapshot_at,selector_snapshot_age_seconds,
older_jobs_excluded,jobs:[.jobs[] | {id,status,stage,sample_count,completed_samples,artifact_available}]}'3. Production setup
The existing Voice Models integration uses trusted sample selection, bounded internal admission and customer-first capacity gates. A new workspace cannot submit through that integration.
Required Core contract
To test your own pipeline, Core must expose a versioned executable recipe with input/output limits and access requirements, tenant-scoped idempotent admission, and durable job/private artifact retrieval. No public recipe for this pipeline currently binds these steps.