{"items":[{"id":"experimental-audio-quality-review","name":"Audio quality comparison","resource_kind":"pipeline","description":"Compare audio quality using specialist predictors and separate listening review.","input":"Source and comparison audio prepared into aligned excerpts.","output":"Separate quality predictions and listening comparisons.","value":"Support comparison of quality dimensions without combining incompatible model scales.","sequence":["prepare aligned source and comparison excerpts","apply specialist quality predictors","compare predictions alongside separate listening review"],"components":[{"model_id":"unilight--sheet-mixed-mos","role":"overall_quality_prediction"},{"model_id":"pam--msclap-2023","role":"clean_quality_prediction"},{"model_id":"blinorot--utmos-2022","role":"naturalness_prediction"},{"model_id":"mustafa-ozan-duman--wavlm-transformer-mos","role":"overall_quality_prediction"},{"model_id":"vkuzgn--scoreq-natural","role":"natural_speech_quality"},{"model_id":"vyvo-research--nisqa-tts","role":"synthetic_speech_naturalness"},{"model_id":"sarulab-speech--utmosv2","role":"naturalness_prediction"},{"model_id":"scoreq--synthetic-nr","role":"synthetic_domain_quality"},{"model_id":"vyvo-research--nisqa-speech","role":"multidimensional_quality_features"},{"model_id":"lj1995--voiceconversionwebui","role":"voice-conversion toolkit reference"},{"inventory_name":"DNSMOS","role":"signal, background, overall and P.808 quality features"}],"availability":{"preview":false,"available":false,"status":"catalog_only"}},{"id":"voice-model-sample-enrichment","name":"Voice sample quality and similarity enrichment","resource_kind":"pipeline","description":"Attach speech-quality features and speaker/audio embeddings to catalog samples for comparison and similarity discovery.","input":"Catalog audio samples prepared according to each component contract.","output":"Quality features, optional gender observations and speaker/audio embeddings.","value":"Support sample comparison and similarity discovery.","sequence":["nisqa","gender","ecapa","clap"],"components":[{"model_id":"vyvo-research--nisqa-speech","role":"quality features"},{"inventory_name":"Gender classifier","role":"optional gender observation"},{"inventory_name":"SpeechBrain ECAPA","role":"speaker embedding"},{"inventory_name":"LAION larger CLAP general","role":"audio embedding"}],"availability":{"preview":false,"available":false,"status":"catalog_only"}},{"id":"private-audio-enrichment-v1","resource_kind":"pipeline","name":"Private audio enrichment","recipe_id":"clap-audio-enrichment-cpu-v1:laion--larger-clap-general:ada0c23a36c4e8582805bb38fec3905903f18b41","alternate_recipes":[{"recipe_id":"clap-audio-enrichment-ivab-v1:laion--larger-clap-general:ada0c23a36c4e8582805bb38fec3905903f18b41","execution_target":"ivab","available":false}],"capability":"audio_enrichment","description":"Four fixed relative audio/text scores and a normalized private CLAP embedding for your own clip.","components":[{"model_id":"laion--larger-clap-general","role":"relative_scores_and_embedding"}],"availability":{"preview":false,"available":false,"status":"execution_disabled"},"input_contract":{"ownership":"caller-owned or explicitly authorized audio only","content_type":"audio/wav","format":"pcm_wav","sample_width_bits":16,"channels":[1,2],"sample_rate_hz":{"min":8000,"max":48000},"min_seconds":2,"max_seconds":10,"max_bytes":6291456,"max_batch_size":1,"max_retained_inputs":32,"privacy":"tenant-private; no catalogue publishing, selection or training","retention":"operator cleanup; no automatic deletion service"},"result_contract":{"schema":"modeluse-audio-enrichment-v1","taxonomy_id":"acoustic-content-four-v1","labels":["a speaking voice","a singing voice","instrumental music","background music"],"embedding_dimensions":512,"confidence":null,"review_required":true,"quality_benchmark":false,"access":"owned GET /v1/artifacts/{artifact_id}; hash-verified private JSON"},"admission_contract":{"endpoint":"POST /v1/jobs","required_body":["pipeline_id","recipe_id","input_id"],"quote":"GET /v1/inputs/{id}/quote?recipe_id=<exact recipe>","meter":"noncash-dev-seconds-v1","units":"ceil(source seconds); delivery once; failure/cancel release","payments_enabled":false,"external_budget_usd":0,"max_outstanding_jobs":20},"execution_contract":{"device":"cpu","activity_slots":1,"shared_with":"whisper-transcription-v1","inference_threads":1,"max_attempts":3,"attempt_wall_seconds":180,"attempt_cpu_seconds":120,"attempt_address_space_bytes":4294967296,"network_downloads":false,"turnaround_sla":null},"qualification":{"model_revision":"ada0c23a36c4e8582805bb38fec3905903f18b41","weight_sha256":"314eb00cce6ad68d25237b8446b659ccdb136ed4672c1bca470f142f72455026","license":"apache-2.0","license_source":"https://huggingface.co/laion/larger_clap_general/blob/ada0c23a36c4e8582805bb38fec3905903f18b41/README.md","evidence_scope":"bounded synthetic CPU execution; no semantic quality guarantee"}},{"id":"whisper-transcription-v1","name":"Private audio transcription","resource_kind":"pipeline","description":"Validate and decode your own WAV, transcribe with pinned Whisper Tiny, and retrieve a verified private result.","input":"One caller-owned or authorized private PCM WAV clip; no catalog selection or URL fetch.","output":"Private transcript, source identity and measured CPU attempt evidence; acoustic classification remains unknown.","value":"Recover a durable transcription result through the same API used for testing and integration.","sequence":["validate owned WAV","decode mono 16 kHz","Whisper Tiny CPU transcription","verify and privately publish result"],"components":[{"model_id":"openai--whisper-tiny","role":"transcription"}],"availability":{"preview":true,"available":true,"status":"qualified_preview"},"recipe_id":"whisper-tiny-cpu-v1:openai--whisper-tiny:169d4a4341b33bc18d8881c4b69c2e104e1cc0af","capability":"transcription","input_contract":{"ownership":"caller-owned or explicitly authorized audio only","privacy":"tenant-private; not published as samples or used for training","content_type":"audio/wav","format":"pcm_wav","sample_width_bits":16,"channels":[1,2],"sample_rate_hz":{"min":8000,"max":48000},"min_seconds":0.1,"max_seconds":30,"max_bytes":6291456,"max_batch_size":1,"max_retained_inputs":32,"retention":"operator cleanup; no automatic deletion service"},"result_contract":{"schema":"modeluse-transcription-v1","fields":["job_id","recipe_id","pipeline_id","input_sha256","input_properties","text","classification","telemetry"],"access":"owned GET /v1/artifacts/{artifact_id}; private hash-verified JSON","classification":"unknown; confidence=null; review_required=true","quality_benchmark":false},"admission_contract":{"endpoint":"POST /v1/jobs","required_body":["pipeline_id","recipe_id","input_id"],"idempotency":"tenant-scoped Idempotency-Key (8–128 characters); identical retries return one job; changed body returns 409","auth":"existing tenant bearer API key or browser session with allowed Origin and X-CSRF-Token for mutations","credits":"existing noncash development credits required; no automatic grants","meter":"noncash-dev-seconds-v1","units":"ceil(source duration_seconds); reserve atomically, settle once on delivery, release on final failure/cancel; retries cost no additional units","max_outstanding_jobs":20,"payments_enabled":false,"external_budget_usd":0},"execution_contract":{"device":"cpu","activity_slots":1,"inference_threads":1,"max_attempts":3,"attempt_wall_seconds":180,"attempt_cpu_seconds":120,"attempt_address_space_bytes":4294967296,"network_downloads":false,"turnaround_sla":null},"qualification":{"model_revision":"169d4a4341b33bc18d8881c4b69c2e104e1cc0af","weight_sha256":"7ebd0e69e78190ffe1438491fa05cc1f5c1aa3a4c4db3bc1723adbb551ea2395","license":"apache-2.0","license_source":"https://huggingface.co/openai/whisper-tiny/blob/169d4a4341b33bc18d8881c4b69c2e104e1cc0af/README.md","evidence_scope":"bounded synthetic CPU execution; no customer quality or turnaround guarantee"}}]}