| Section | Description |
|---|---|
| Settings | Object-specific settings that change runtime behavior. |
| Dialing | Settings shown in the Dialing block. |
| Diversion field for External Agents | Agent assignment and queue membership behavior. |
| Autocreated Custom Destination | Settings shown in the Autocreated Custom Destination block. |
| Recording | Recording storage and recording-related options. |
| Voice Synthesizer | Voice calling behavior, caller handling, recording, and voice-specific limits. |
| Speech to Text | Settings shown in the Speech to Text block. |
| Generative Artificial Intelligent chat models | Settings shown in the Generative Artificial Intelligent chat models block. |
| Dial By Name Directory | Settings shown in the Dial By Name Directory block. |
| CNAM service | Settings shown in the CNAM service block. |
| Faxing | Fax handling, quality, storage, delivery, and fax-specific transport settings. |
| MS Teams integration | Settings shown in the MS Teams integration block. |
| API Interface | API-related credentials, restrictions, and integration controls. |
| Default Music On Hold | Settings shown in the Default Music On Hold block. |
| Unassigned DID destination | Settings shown in the Unassigned DID destination block. |
| General DID destination | Settings shown in the General DID destination block. |
| General Outbound destination | Settings shown in the General Outbound destination block. |
| Emergency numbers notification | Settings shown in the Emergency numbers notification block. |
| Security | Tenant security, limits, and access controls. |
| Call Settings | Object-specific settings that change runtime behavior. |
| Outbound Calls | Settings shown in the Outbound Calls block. |
| CSV Exports | Settings shown in the CSV Exports block. |
| Voicemail Settings | Voice calling behavior, caller handling, recording, and voice-specific limits. |
| Voicemail backup | Voice calling behavior, caller handling, recording, and voice-specific limits. |
| Field | Meaning |
|---|---|
| Engine | Use **System Default** to inherit the global Speech to Text engine. This is the normal choice when all tenants share the same transcription provider. |
| Language | Leave as **System Default** to inherit the global language selection. If a tenant override is available, select the language expected for that tenant audio. |
| Speech To Text Engine API password | Leave empty to inherit the global secret. Fill it only when the tenant must override the credential used by the inherited engine. |
| Minimum call time for transcript | Calls shorter than this number of seconds are not sent for transcription. This avoids sending very short recordings to the external service. |
| Test speech recognition | Upload a WAV file and select **Test** to verify the currently selected engine configuration before relying on it for call transcripts. |
| Field | Meaning |
|---|---|
| Engine | Selects IBM Watson as the tenant Speech to Text provider. The historical engine name includes TTS, but this block is used for speech recognition. |
| Full Endpoint URL | Watson recognition endpoint, including the service instance path. Use the recognize endpoint supplied by IBM for the tenant account or service instance. |
| Language | Recognition model used by Watson, such as an American English broadband model. Choose the model that matches the audio language and expected bandwidth. |
| Speech To Text Engine API username | Watson credential username. Many IBM deployments use `apikey` as the username when authenticating with an API key. |
| Speech To Text Engine API password | Watson API key or password. The screenshot uses a fictional value. |
| Minimum call time for transcript | Minimum recording duration, in seconds, before MiRTA PBX attempts transcription. |
| Test speech recognition | Runs a test upload through the configured Watson settings. |
| Field | Meaning |
|---|---|
| Engine | Selects Nexiwave as the tenant Speech to Text provider. |
| Language | Language sent to Nexiwave for recognition. The installed option shown in the example is English. |
| Speech To Text Engine API username | Nexiwave account username for the tenant-specific integration. |
| Speech To Text Engine API password | Nexiwave password or API secret. The screenshot uses a fictional value. |
| Minimum call time for transcript | Minimum recording duration before the call is eligible for transcription. |
| Test speech recognition | Uploads a WAV file to test the configured Nexiwave credentials and language. |
| Field | Meaning |
|---|---|
| Engine | Selects Google Cloud Speech to Text for tenant transcription. |
| Language | Google recognition language code. Choose the language that best matches the callers; the example uses English (United States). |
| Speech To Text Engine API password | Google API key used when audio is submitted directly without a storage bucket. Leave empty only when the deployment inherits the value elsewhere. |
| JSON Service Account key | Service account JSON used when audio is uploaded to Google Cloud Storage before transcription. The account must have access to the bucket. |
| Bucket Name | Google Cloud Storage bucket used to stage audio for transcription. The bucket must already exist and be accessible by the service account. |
| Profanity filter | Controls whether Google should attempt to mask profanity in returned transcripts. Use **System Default** to inherit the global behavior, or force **Yes** or **No** for this tenant. |
| Minimum call time for transcript | Minimum recording duration before MiRTA PBX sends the call to Google. |
| Test speech recognition | Uploads a WAV file and validates the Google configuration. |
| Field | Meaning |
|---|---|
| Engine | Selects AssemblyAI as the tenant Speech to Text provider. |
| Endpoint server | AssemblyAI API host. Leave empty for the platform default, or set a host such as `api.assemblyai.com` when documenting or overriding the endpoint. |
| Model | AssemblyAI model used for the transcript request. Use the stronger model when accuracy matters, or a fallback combination when availability is more important. |
| Language | Language sent to AssemblyAI. **Automatic** lets AssemblyAI detect the language, while a specific language restricts recognition to that language. |
| Speech To Text Engine API password | AssemblyAI API token. The screenshot uses a fictional value. |
| Minimum call time for transcript | Minimum recording duration before the call is eligible for transcription. |
| Test speech recognition | Uploads a WAV file to test the AssemblyAI endpoint, model, token, and language choices. |
| Field | Meaning |
|---|---|
| Engine | Selects GeminiAI for speech recognition through a generative AI model. |
| Model | Gemini model used for transcription. Select the model according to the required balance of speed, cost, and accuracy. |
| Model command | Instruction sent to the model for the transcription task. Leave it empty to use the default command, or provide tenant-specific instructions such as speaker separation and terminology handling. |
| Language | Recognition language. The GeminiAI language list can be left at the inherited value or set to an available language option such as any language. |
| Speech To Text Engine API password | Gemini API key. The screenshot uses a fictional value. |
| Minimum call time for transcript | Minimum recording duration before MiRTA PBX asks GeminiAI to transcribe the call. |
| Test speech recognition | Uploads a WAV file and verifies the GeminiAI model, command, key, and language settings. |
| Task | How to do it |
|---|---|
| Edit | Change the required tenant settings in the relevant block, then select Save. |
| Linked pages | Use the buttons inside Settings to open related management pages such as CNAM cache, manager users, routing rules, and PIN codes. |
| Delete | Settings itself is not deleted. Delete actions are handled on the linked object pages. |