Overview
Twilio offers transcription services primarily through its "Conversational Intelligence (Classic)" suite, which provides modern, accurate models for both real-time and offline use cases.
It works in two steps:
- Transcription: It converts voice audio into text, either in real-time (streaming) or from a recording (batch).
- Analysis: It then uses AI models called Language Operators to analyze that transcript for insights like sentiment, PII redaction, summarization, and custom business logic.
You pay for both the transcription and the analysis that runs on top of it.
What You Need To Know
Real-Time Transcription (Streaming)
This service provides low-latency transcripts as the call is happening, allowing for live insights and agent-assist applications.
- Service: Conversational Intelligence - Streaming (Real-Time) Transcription
- Use Cases: Live sentiment analysis, real-time agent coaching, live captioning, and triggering in-call automations.
- Price: $0.027 / min
Offline Transcription (Batch)
This service is used to transcribe audio from existing recordings after a call is completed.
- Service: Conversational Intelligence - Batch Transcription
- Use Cases: Post-call quality assurance, compliance audits, customer feedback analysis, and training AI models.
- Price: $0.024 / min
Transcription with Analysis (Language Operators) Pricing
When you use Conversational Intelligence, your total cost is typically (A) Transcription Price + (B) Analysis Price. The analysis pricing depends on the type of AI model (Language Operator) you use.
Here is the full pricing breakdown:
| Service Component | Price | Description |
| A: TRANSCRIPTION (Choose one) | ||
| Batch Transcription (Offline) | $0.024 / min | Transcribes existing audio recordings. |
| Streaming Transcription (Real-Time) | $0.027 / min | Transcribes audio live as it happens. |
| B: ANALYSIS (Language Operators) | ||
| Standard Operators | $0.0040 / min | For pre-built models (e.g., sentiment analysis, PII redaction). |
| Text Analysis | $0.0035 / min | For other standard text analysis tasks. |
| Generative Custom Operators |
Input: $0.002 / 1k chars Output: $0.018 / 1k chars |
Uses LLMs for custom, generative tasks (e.g., "Was the agent empathetic?" or "Score this call for compliance."). |
Note on "Standard" Transcription: It's also worth noting that Twilio offers a basic transcription service as part of its standard Call Recording feature. This service is priced at $0.0500 / min.
Reference: Transcribe entire phone calls with Twilio
Real-Time vs. Offline Summary
| Feature | Real-Time (Streaming) | Offline (Batch) |
| Use Case | Live agent assist, real-time alerts | Post-call analysis, QA, compliance |
| Transcription Price | $0.027 / min | $0.024 / min |
| Analysis Price | + Analysis Operator cost (e.g., $0.0040/min) | + Analysis Operator cost (e.g., $0.0040/min) |
| Total Example Cost | ~$0.031 / min (Transcription + Standard Operator) | ~$0.028 / min (Transcription + Standard Operator) |
Pricing is subject to change, so it's best to check the Twilio website for the most recent price: Twilio Voice Pricing
Additional Information