Process a conversation
To generate conversation intelligence and pre-built UI experiences for you, you must provide the Symbl.ai APIs with a conversation. Conversations come in many forms:
- Meetings in your organization.
- Calls with customers.
- Transcripts and recordings of client interactions.
- Chat and email threads.
The Symbl.ai APIs provide multiple ways to process conversations. The following table provides a quick breakdown of the APIs for processing conversations.
| API | Description |
|---|---|
| Async API | The Async API processes conversation intelligence from previously-recorded text, audio, and video. Symbl.ai provides the following methods of processing conversations with the Async API: |
| Streaming API | The Streaming API supports the WebSocket protocol to process audio and generate real-time intelligence from live conversations. |
| Telephony API | The Telephony API supports the PSTN and SIP protocols and allows Symbl.ai to bridge or join VoIP calls and generate real-time conversation intelligence. |
After you process a conversation, Symbl.ai provides additional APIs for generating Conversation Intelligence.
Conversation ID
When you process a conversation, whether real-time or asynchronously, the Symbl.ai APIs return a conversation ID. The conversation ID is the unique identifier of a conversation and a critical component of your API requests when you generate conversation intelligence.
When you make a request to the Symbl.ai APIs for intelligence, you use the conversation ID to target a specific conversation. You can also use conversation IDs to assemble a conversation from multiple sources, such as two or more videos from a series of meetings.
Async API
The Async API provides paths for processing conversations in recorded video, audio, and text. The Async API also lets you collect multiple audio and text sources into a single conversation.
Features of the Async API:
- Process conversations from video (.mp4), audio (.m4a, .mp3, .wav), and text.
- Detect actionable phrases and entities such as organizations, dates, people, and more.
- Identify individual speakers in multichannel audio.
- Provide transcripts of audio in many languages.
- Implement trackers to detect custom phrases and keywords.
- Get AI-generated summaries of your conversations.
- Score a conversation with qualitative feedback.
To get started asynchronously processing your conversations, see Async API.
Streaming API
The Streaming API supports the WebSocket protocol to process audio and generate in-the-moment intelligence from real-time conversations.
To get started processing your WebSocket-based real-time conversations, see Streaming API.
Telephony API
The Telephony API supports SIP and PSTN and allows Symbl.ai to bridge or join VoIP calls and generate real-time conversation intelligence.
To get started processing your SIP-based real-time conversations, see Telephony API.
Updated 5 months ago