Help and limitations
What IVRReady validates
IVRReady preflights and packages IVR audio for contact-center platforms. It is an independent third-party validator using each platform’s published audio file requirements; it is not affiliated with, endorsed by, or certified by any vendor. It checks documented file-format constraints; it does not verify prompt wording, language, contact-flow behavior, platform permissions, tenant configuration, or subjective voice quality.
- Amazon Connect
- Accepted upload: RIFF/WAVE
- Codec: G.711 μ-law (WAV format ID 0x0007)
- Sample rate: 8,000 Hz
- Channels: Mono
- Duration: strictly less than 300 seconds
- File size: strictly less than 50 MB (decimal)
- Webex Contact Center
- Accepted upload: RIFF/WAVE
- Codec: G.711 μ-law (8-bit u-law) (WAV format ID 0x0007)
- Sample rate: 8,000 Hz
- Channels: Mono
- Sample size: 8-bit
- Duration: no generic limit documented by the platform
- File size: no generic limit documented by the platform
- NICE CXone
- Accepted upload: RIFF/WAVE
- Codec: CCITT μ-Law (G.711) (WAV format ID 0x0007)
- Sample rate: 8,000 Hz
- Channels: Mono
- Sample size: 8-bit
- Bit rate: 64 kbps
- Duration: no generic limit documented by the platform
- File size: no generic limit documented by the platform
- Genesys Cloud Architect
- Accepted upload: WAV
- Accepted formats: 10 documented WAV combinations
- Sample rate: 8,000 or 11,025 Hz
- Channels: Not stated in the cited supported-format matrix
- Sample size: 8-bit
- Duration: no generic limit documented by the platform
- File size: 25 MB maximum; IVRReady uses 25,000,000 bytes
- Runtime conversion: Architect converts uploaded prompt audio to 8-bit μ-law for runtime playback.
- Talkdesk Greetings
- Accepted upload: MP3, WAV
- IVRReady coverage: WAV inputs; Talkdesk MP3 intake is not implemented
- Sample rate: not constrained for WAV upload; runtime format is 8,000 Hz
- Channels: Not stated as a greeting-upload requirement on the cited page; runtime format is mono
- Sample size: 8-bit
- Duration: no generic limit documented by the platform
- File size: Talkdesk guidance says not to exceed 10 MB; no hard generic greeting limit is established by the cited page
- Runtime conversion: Talkdesk says any uploaded audio is transcoded to 8-bit mono μ-law at 8,000 Hz.
- Twilio Media Streams
- Scope: outbound audio in a bidirectional <Connect><Stream>
- IVRReady source input: WAV
- Delivered media: audio/x-mulaw
- Sample rate: 8,000 Hz
- Channels: mono
- Transport payload: base64
- Container header: none; raw audio bytes only
- Payload size: Twilio documents any size; practical browser memory and IVRReady safety limits still apply
Four different support boundaries
- Structurally recognized
- The RIFF/WAVE parser can identify PCM, float, G.711 A-law/μ-law, IMA ADPCM, MP3-in-WAV, and safely formed extensible WAV metadata. Recognition alone is not readiness.
- QA-decodable
- Audio QA decodes PCM 8/16/24/32-bit, float32, A-law, and μ-law. IMA ADPCM and MP3-in-WAV are structurally recognized but currently show QA unsupported.
- Automatically correctable
- Correction is limited to the source layouts verified with the pinned real FFmpeg/WASM engine. Unlisted rates, channel layouts, float64, IMA ADPCM, and MP3-in-WAV are not advertised as correctable merely because FFmpeg may recognize them.
- PCM integer: 8-bit 8 kHz mono; 16-bit 8 kHz mono, 44.1/48 kHz stereo; 24/32-bit 48 kHz stereo
- IEEE float32: 8 kHz mono and 48 kHz stereo
- G.711 A-law: 8 kHz mono
- G.711 μ-law: 8 kHz mono/stereo and 16 kHz mono
- WAVE_FORMAT_EXTENSIBLE: 24-bit PCM and float32 at 48 kHz stereo with the verified GUID/mask layouts
- Target-ready
- The source or generated bytes pass every IVRReady format check listed above for the selected target pack.
Statuses and QA
Ready means compatibility passed, source QA completed, and no QA warning exists. Fixable means only verified mechanical corrections are required. Warning means compatibility passed but QA found a policy warning or was unavailable. Error means a blocking format or structural failure. QA measures peak, possible repeated digital-limit samples, signal-level silence, and edge silence; it is not speech recognition, intelligibility scoring, or clipping proof.
Local build and verification
Selected audio and file-associated metadata stay in the browser media workflow. Already-ready prompts are copied byte-for-byte. Corrected output is re-inspected structurally from generated bytes before packaging. For Twilio Media Streams, the proven canonical WAV is instead materialized as raw μ-law bytes and a base64 media-message template. Source audio receives QA; corrected output receives structural target validation, not a second audio-QA pass.
ZIP and batch limits
ZIP intake rejects traversal, encrypted or multi-disk archives, ZIP64, CRC failures, unsafe collisions, entries over 100 MB, total expansion over 1 GB, ratios over 100:1, and ZIP files over 500 MB. The 2,000-prompt limit is a safety ceiling, not a routine performance promise. Up to 500 typical short prompts was the tested recommendation on the documented desktop Chromium environment.
Browser and output limitations
The complete workflow was validated on Playwright Chromium and WebKit and on the tested branded Chrome and Edge versions. Playwright WebKit is not Safari. Actual Safari remains untested. Firefox remains unvalidated because the current Windows test host could not launch its compositor; that is recorded as not tested, not unsupported.
ZIP download is the universal output workflow. Save to folder is a browser-dependent File System Access enhancement and appears only where the browser exposes it. Native operating- system picker behavior still requires manual browser validation.