Space supports output streaming and WebSocket input streaming. Not all access methods have the same latency contract.
Batch processing
Not collected
No model-specific specification verified; unknown is not unsupported.
Fine-tuning
Not collected
No model-specific specification verified; unknown is not unsupported.
Reasoning
Not collected
No model-specific specification verified; unknown is not unsupported.
Context caching
Not collected
No model-specific specification verified; unknown is not unsupported.
Base model
Not collected
No model-specific specification verified; unknown is not unsupported.
Creation privacy
Website policy retains personal data as necessary, with legal exceptions and international transfers. Its separate AI Voice Assistant appendix promises no model training or voice cloning for assistant conversations and specifies 12-month transcript/technical-data retention. Those assistant-only terms are not a Space TTS or Marketplace inference-retention guarantee.Source for Creation privacy (opens in a new tab)
Provider calls each language base URL a model. SDK predefined environments cover only English and Ukrainian; the broader language table supplies the other three URLs. API v1 is not a checkpoint release.
Maximum prompt characters
Not collected
Bytes recommends approximately 5,000 characters under a 120-second request timeout; longer input uses streaming. No universal per-model character ceiling established.
Authenticated GET /voices at the chosen model base URL returns IDs, optional full name, gender, accent, age and per-voice sampling defaults. Samantha and Amara are documentation examples, not a complete per-language inventory.Source for Stock voice catalog (opens in a new tab)
Voice controls
Voice-specific defaults may be overridden: temperature >= 0; top_k = -1 or > 0; top_p > 0 and <= 1; min_p 0–1; presence/frequency penalties 0–2; repetition penalty 1–2; optional integer seed. Provider promises identical generation for identical parameters with a non-null seed; this was not independently tested and is not caching.Source for Voice controls (opens in a new tab)
Bytes WAV contract is separately sourced from the Bytes reference; transport encodings are not different models.
Published price
USD 2 per 60,000 characters for current Space catalogue voices; pay as you go with top-up/auto-recharge. The approximate USD 2/hour illustration is not a separate hourly charge. Custom voices and volumes require a quote.Source for Published price (opens in a new tab)
Documentation discrepancy
HTML examples label a 204 response; Markdown specifies HTTP 200 and the official SDK reads binary WAV or JSONL. Do not parse Bytes as JSON despite generated sample snippets. SDK/server environment lists cover two languages while the model catalogue lists five.Source for Documentation discrepancy (opens in a new tab)
Response contract corroborated by official SDK and separate model-language reference; inference not tested.
Access & pricing
Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.
Voice-specific defaults may be overridden: temperature >= 0; top_k = -1 or > 0; top_p > 0 and <= 1; min_p 0–1; presence/frequency penalties 0–2; repetition penalty 1–2; optional integer seed. Provider promises identical generation for identical parameters with a non-null seed; this was not independently tested and is not caching.Source for Voice controls (opens in a new tab)
Approximately 5,000 characters under a 120-second HTTP timeout. Bytes response cap is 100 MB and is silently truncated when exceeded; use streaming for longer input.Source for Input limits (opens in a new tab)
Prices retain their published currencies and billing units below.
Published billing units · original currency · observed 16 Oct 2026
Category
Usage
Rate
Unit
Text-to-speech
Space catalogue voices1
$2.00
60,000 characters
Rate conditions
Same Space tariff across alternative access methods, not additive charges. Character billing; approximate audio-hour illustration is not a token or question price. Custom arrangements may differ.
Service-specific terms and exact inference retention require confirmation. Website policy retains personal data as necessary, with legal exceptions and international transfers. Its separate AI Voice Assistant appendix promises no model training or voice cloning for assistant conversations and specifies 12-month transcript/technical-data retention. Those assistant-only terms are not a Space TTS or Marketplace inference-retention guarantee.
Language-specific route selector, not an immutable checkpoint ID.
Access
POST https://api.respeecher.com/v1/public/tts/ar-rt/tts/sse. Whole transcript input; output is JSON-lines chunk/error objects with base64 raw PCM, despite the SSE endpoint name.
Regions
Inference residency not verified; company location is not processing-region evidence.
Voice-specific defaults may be overridden: temperature >= 0; top_k = -1 or > 0; top_p > 0 and <= 1; min_p 0–1; presence/frequency penalties 0–2; repetition penalty 1–2; optional integer seed. Provider promises identical generation for identical parameters with a non-null seed; this was not independently tested and is not caching.Source for Voice controls (opens in a new tab)
Prices retain their published currencies and billing units below.
Published billing units · original currency · observed 16 Oct 2026
Category
Usage
Rate
Unit
Text-to-speech
Space catalogue voices1
$2.00
60,000 characters
Rate conditions
Same Space tariff across alternative access methods, not additive charges. Character billing; approximate audio-hour illustration is not a token or question price. Custom arrangements may differ.
Service-specific terms and exact inference retention require confirmation. Website policy retains personal data as necessary, with legal exceptions and international transfers. Its separate AI Voice Assistant appendix promises no model training or voice cloning for assistant conversations and specifies 12-month transcript/technical-data retention. Those assistant-only terms are not a Space TTS or Marketplace inference-retention guarantee.
Language-specific route selector, not an immutable checkpoint ID.
Access
GET wss://api.respeecher.com/v1/public/tts/ar-rt/tts/websocket with X-API-Key handshake. Send transcript, voice and context_id; continue=true preserves prosody across all but the final text chunk.
Regions
Inference residency not verified; company location is not processing-region evidence.
Voice-specific defaults may be overridden: temperature >= 0; top_k = -1 or > 0; top_p > 0 and <= 1; min_p 0–1; presence/frequency penalties 0–2; repetition penalty 1–2; optional integer seed. Provider promises identical generation for identical parameters with a non-null seed; this was not independently tested and is not caching.Source for Voice controls (opens in a new tab)
Responses are chunk, done or error, mapped by context_id. Cancel with context_id and cancel=true; audio already generated but not yet streamed is still charged. Connections are reusable, but lifetime beyond 12 hours is not guaranteed and idle connections can close.Source for Response format (opens in a new tab)
Connection lifetime comes from the linked WebSockets guide. Protocol JSON is not caller-defined structured generation.
Prices retain their published currencies and billing units below.
Published billing units · original currency · observed 16 Oct 2026
Category
Usage
Rate
Unit
Text-to-speech
Space catalogue voices1
$2.00
60,000 characters
Rate conditions
Same Space tariff across alternative access methods, not additive charges. Character billing; approximate audio-hour illustration is not a token or question price. Custom arrangements may differ.
Service-specific terms and exact inference retention require confirmation. Website policy retains personal data as necessary, with legal exceptions and international transfers. Its separate AI Voice Assistant appendix promises no model training or voice cloning for assistant conversations and specifies 12-month transcript/technical-data retention. Those assistant-only terms are not a Space TTS or Marketplace inference-retention guarantee.
Caching
Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.
01 / INITIAL INPUTProcess & retain
The system processes eligible input and retains reusable state.
02 / LATER REQUESTCheck available state
Reuse depends on matching input, session state and the cache policy.
03 / CACHE HITReuse & generate
Reusing state can reduce repeated work. Any charges follow the billing rules below.
Respeecher Space API
ar-rt
Not verified
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
How it works
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Minimum reusable input
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Lifetime & refresh
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
How to enable it
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Usage fields
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Billing rules
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Limits & data boundaries
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
No verified cache rate for this endpoint. An empty rate is not a zero charge.
Caching reviewed 16 Oct 2026
Respeecher Space API
ar-rt
Not verified
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
How it works
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Minimum reusable input
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Lifetime & refresh
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
How to enable it
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Usage fields
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Billing rules
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Limits & data boundaries
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
No verified cache rate for this endpoint. An empty rate is not a zero charge.
Caching reviewed 16 Oct 2026
Respeecher Space API
ar-rt
Not verified
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
How it works
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Minimum reusable input
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Lifetime & refresh
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
How to enable it
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Usage fields
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Billing rules
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
Limits & data boundaries
Not collected
No inference-cache contract verified. WebSocket connection reuse, continuing prosody context, deterministic seeds and saved recordings do not establish discounted input caching.
No verified cache rate for this endpoint. An empty rate is not a zero charge.
Caching reviewed 16 Oct 2026
What the measurements say
No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.
Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.
The record behind the model
Why it is included
Provider describes proprietary deep-learning speech technology and identifies this model family in its own documentation. Public client libraries are not model weights; individual speakers and API versions are not separate checkpoints.
