We often encounter the situation where call recording is enabled for everyone, but no one manually listens to hour-long conversations. That's where our transcription setup turns audio into searchable, analyzable text and pushes it into CRM without listening. We offer a turnkey solution: from cloud activation to on-premise integration via external Speech-to-Text services. We assess your project in one day and complete the setup within 2-5 business days. Our extensive experience with Bitrix24 and dozens of implemented projects ensure stable recognition even with suboptimal recording quality.
How call transcription actually works
Bitrix24 uses the built-in Bitrix24 Contact Center module together with a cloud speech recognition service. After a call ends, the recording is sent via the telephony module to the recognition server — the result is returned as text and saved to the CRM entity card (lead, deal, contact).
Technically, the process looks like this:
- The call ends, the recording is saved in
b_voip_calllinked toCALL_ID. - The transcription task enters the handler queue.
- The Speech-to-Text service (Yandex SpeechKit or the built-in Bitrix engine) receives the audio file.
- The text is returned and written to the
TRANSCRIPTfield of the call record. - A "Call transcription" block with speaker separation appears in the CRM card.
What is required for activation
Transcription is available only on Professional and Enterprise plans (Bitrix24 cloud). For on-premise, you need an external Speech-to-Text provider and the REST API method telephony.externalcall.finish.
Activation steps:
- CRM → Telephony → Settings → "Transcription" tab.
- Enable the toggle "Transcribe calls".
- Select recognition language (Russian, English — list depends on plan).
- Specify which phone numbers to enable transcription for.
- Set up access rights: who sees transcriptions (manager, employee themselves, everyone).
After activation, every new call longer than 10 seconds will be automatically transcribed. Old recordings — only via manual trigger or REST script.
Limitations and common issues
Recognition quality strongly depends on recording quality. If the line has noise, echo, or low bitrate, accuracy drops to 60-70%. Minimum recommended bitrate: 8 kHz mono, preferably 16 kHz.
Speaker separation (diarization) works only with stereo recordings where operator and client are on separate channels. Mono recordings produce a single text block without speaker labels.
Quotas. Each plan has a monthly limit of transcription minutes. When exceeded, transcription stops until the next billing period. Limits are visible under "Tariffs" → "Usage".
| Deployment | Provider | Diarization | Minute limit | License cost |
|---|---|---|---|---|
| Cloud (Pro/Enterprise) | Built-in engine | Yes (stereo) | Included in plan | Included |
| On-premise | Yandex SpeechKit / Google STT | Yes (stereo) | Unlimited (pay per character) | Separate subscription |
On-premise version. The built-in engine is not available. Integration with Yandex SpeechKit or Google Speech-to-Text is implemented via the OnVoipCallEnd event handler, calling an external API and writing the result back to Bitrix via CRest::call('telephony.externalcall.finish', [...]).
What to do with the transcribed text
Transcripts can be used in robots and business processes: trigger "Call completed" → condition on transcript text → action (change stage, assign task, notify). The transcript field is available as the {=Document.TRANSCRIPT} variable in the business process designer.
Search by transcript texts works via standard CRM search — the transcript is indexed in b_search_content along with other fields. Time savings for call analysis reach 70%.
How to improve recognition accuracy to 95%?
Accuracy directly depends on three factors: recording bitrate (recommended 16 kHz mono or stereo), noise level (noise suppression on telephony), and speech clarity. With stereo recording on separate channels and 16 kHz bitrate, accuracy exceeds 90%. Mono recordings with noise drop to 60-70%. We recommend auditing your current telephony before setup. Read more about quality parameters in Bitrix24 documentation.
What is included in transcription setup
| Stage | Result | Effort |
|---|---|---|
| Audit of current telephony and CRM | Integration scheme, list of required accesses | 1-2 hrs |
| Activation of transcription on the cloud | Working transcription on selected numbers | 1-3 hrs |
| Setting rights and filters by number | Access control for transcriptions | 1-3 hrs |
| On-premise integration via SpeechKit | Stable recognition with diarization | 8-16 hrs |
| Configuration of robots and business processes | Automatic actions based on transcript text | 2-8 hrs |
| Documentation and training | Admin manual, manager training | 1-2 hrs |
Pricing & deliverables
Typical project cost: $500–$2000 depending on complexity. Includes 3 months of post-setup support. Our team has 10+ years of experience in Bitrix24 integration and over 50 completed transcription projects, guaranteeing quality and fast turnaround.Have questions? Contact us — we will assess your project for free and recommend the optimal solution. Order call transcription setup in Bitrix24 and get a fully automated process for conversation analysis. With guaranteed recognition quality and certified expertise, you save up to 40 hours per month per manager — equivalent to $800+ in productivity gains.