Firetell
Feature Overview

Live Transcription for active phone calls

Convert speech into text while the call is still in progress with < 300ms latency. Stream live text to agent portals, translate subtitles on the fly, and push structured transcripts to your CRM.

Real-time streaming with < 300ms latency
Dual-channel stereo separation (Client vs Agent)
Unique Dialogue IDs for Live Translation
Automatic Webhook dispatch & AI Summary ready
Capturing Live

Speaker Timeline

Live transcript feed

Streaming audio

Call ID

FT-240391

Latency

< 300ms

Detected Intent

Order status check

Agent

Keep the agent channel isolated for QA reviews, coaching, and compliance checks.

Client

Capture customer speech clearly with dedicated audio channel separation for real-time translation and CRM lookup.

Example

00:00 - 00:22

Agent

dlg_1
99%

Hello, thank you for calling. How can I help you today?

Client

dlg_2
99%

Hi, I'd like to check my order status, my reference code is FT10294.

Agent

dlg_3
99%

Got it. Please hold on for just a second while I pull up your order information.

Client

dlg_4
99%

Sure, I want to make sure it was shipped out since I haven't received the tracking email.

Transcribing live speech stream...

Live Insight

The system can use the live transcript to suggest intent, trigger real-time translation subtitles, or push CRM context while the call is still active.

Core Capabilities

Focus on the conversation while it is happening

This page introduces how Firetell presents transcripts in a clear, readable format that is ready for contact center operations.

Ultra-low latency streaming

Powered by advanced real-time Speech AI models, speech is converted into text in under 300ms so agents and systems react immediately.

Stereo dual-channel separation

Cleanly isolates Client and Agent audio on separate legs, ensuring pristine speaker tagging without crosstalk.

Webhook & CRM Integration

Automatically dispatches 'call.transcription.completed' events with structured dialogue turns directly into your CRM or data warehouse.

How It Works

From live audio to structured transcript

This page outlines a typical flow so the transcript is not only readable on screen, but useful for operations and analytics.

1

Capture live call audio

Firetell streams 16kHz linear PCM audio over WebSockets directly to the AI Gateway while the call is active.

2

Separate channels and tag speakers

The system distinguishes Client and Agent legs with unique dialogue IDs, making real-time translation seamless.

3

Deliver transcript into operations

Transcripts stream to agent portals in real-time, persist in dedicated collections upon hangup, and dispatch to webhooks.

Structured Output

More than text, it is usable operational data

When transcripts are structured from the start, teams can search, analyze, and automate much faster.

Speaker Labels

Each utterance is tagged as Agent or Client with unique dialogue IDs for reliable timeline mapping.

Timeline & Confidence

Includes precise millisecond timestamps, speech duration, and confidence scores per dialogue segment.

Searchable Full-text

Full-text indexed records make it easy to search for keywords, order IDs, issues, or intents across all calls in seconds.

AI Insights Ready

Pre-structured for post-call AI summaries, sentiment scoring, action items checklists, and QA evaluations.

Developer First

Instant integration with REST APIs & Webhooks

Consume streaming SSE events during the call or subscribe to the 'call.transcription.completed' webhook to sync transcripts directly into your CRM or data warehouse.

POST /your-webhook-endpoint HTTP/1.1
Content-Type: application/json
X-Firetell-Signature: sha256=a8f4c2e...

{
  "event": "call.transcription.completed",
  "workspace_id": "ws_01h900000000000000000001",
  "timestamp": "2026-08-18T07:00:22.000Z",
  "data": {
    "id": "trans_01h900000000000000000001",
    "call_id": "call_01h900000000000000000001",
    "direction": "inbound",
    "dialogue_count": 4,
    "full_text": "Client: Hello...\nAgent: Certainly...",
    "action_items": [
      "Send package tracking link to client email"
    ]
  }
}
Operational Value

Clear benefits for support, QA, and supervisors

When conversations become data during the call itself, many downstream manual tasks can be reduced.

Find the right conversation segment faster instead of replaying the full recording.
Support audits and compliance checks with clearly labeled Client and Agent speaker turns.
Enable real-time alerts when important keywords or escalation signals appear.
Create a strong foundation for analytics dashboards and call quality reporting.
Reduce post-call review time for operations and QA teams.
Sync automatically with CRMs (HubSpot, Salesforce, Zoho) via Webhook events.
Use Cases

Where does Live Transcription fit best?

This feature is most valuable when transcripts need to be used during operations, not just stored for later playback.

Contact Center Monitoring

Supervisors can follow live calls without having to listen continuously to the full audio stream.

Real-time Translation & Subtitles

Agents can view translated subtitles instantly for foreign-language callers using unique dialogue turn IDs.

Compliance & QA Review

Quality and legal teams can review required disclosures or sensitive wording much more efficiently.

AI Assist & Summaries

Realtime transcripts provide the foundation for summaries, action items, suggested replies, and CRM syncing.

Pricing

Simple pricing for realtime transcription

Start with a clear usage-based model that is easy to forecast and flexible enough for both pilots and growing contact center workloads.

Pay as you use

Live Call Transcription

One straightforward rate for live transcription, with no upfront commitment and billing based only on active call minutes.

$0.01/min

Billed per minute directly from workspace credit balance

Included

  • Next-Gen Speech AI Engine
  • Multi-language & Auto-detect
  • Dual-channel Stereo Separation
  • Real-time Timestamps & IDs
  • Confidence & Word-level Scores
  • Automatic Webhook & CRM Sync

Pricing Examples

1,000 mins / month

$10

10,000 mins / month

$100

100,000 mins / month

$1,000

Usage-Based

Easy to start, no complex packages

A good fit for teams that want to launch quickly or scale gradually with real traffic instead of committing to a fixed bundle upfront.

Production Ready

Rich enough for real operational workflows

Capabilities like channel separation, timestamps, and confidence scores make the transcript useful for QA, analytics, and CRM automation.

Best Fit

Ideal when transcript visibility matters during the call

This works especially well for live monitoring, real-time translation, supervisor assist, or any workflow that needs call content in real time.

Toggle in Console

Live Transcription can be enabled or disabled per phone number directly in the console with 1-click.

Interactive Cost Estimator

Estimate your monthly transcription cost

Drag the slider to calculate how much you will spend based on your expected monthly active call volume.

Monthly Cost

$5.00

Estimated Monthly Call Minutes500 mins / mo
500 mins25,000 mins50,000 mins100,000 mins

Flat $0.01 / minute • No minimum commitments • Pay per second used

Ready to Deploy

Bring live transcription into your calling workflow

If you want to combine transcription with recording, analytics, or Voice AI, Firetell can help design the right rollout.