Audio reference and metadata accepted.
Conversation intelligence API for product teams.
Create conversation objects, retrieve source moments, run agents, and receive webhooks when conversation intelligence needs to live inside your product or tools.
Agent output can trigger an application workflow.
conversation records
over conversation objects
workflow events
{
"conversation_id": "conv_9j2f7",
"agent": "custom_escalation_check",
"source_moment": {
"timestamp": "00:42",
"speaker": "Workflow owner",
"signal": "Webhook candidate"
},
"destinations": ["Object","Agent run","Webhook"],
"webhook": "agent_run.completed"
}Cited result available for review.
Destination received the mapped payload.
Async by design. Structured by default.
The API is built around queued processing, reusable conversation objects, agent runs, and webhook delivery.
POST /conversations
Send audio, transcript references, metadata, source system, and custom fields. Receive a conversation ID immediately.
GET /conversations/:id
Retrieve status, segments, speakers, timestamps, metadata, processing state, and object fields.
POST /agent-runs
Run Sales QC, analytics, customer, event, or custom checks over a conversation or conversation set.
Receive webhooks
Handle conversation.created, transcription.completed, agent_run.completed, and workflow.triggered events.
Build on the object, not raw transcripts.
Product teams use Speechrun when they need source moments, agent outputs, and workflow payloads inside their own software.
Conversation object
Segments, speakers, timestamps, metadata, custom fields, and processing state.
Agent result
Structured finding, source moment, severity, reason, destination, and review context.
Webhook payload
Event type, IDs, status, source moment reference, and application route context.
Search and retrieval
Find commitments, risks, objections, customer signals, event themes, or custom tags across conversations.
Questions about API access.
How conversation objects, agent runs, and webhooks fit into a workflow.
Map the API workflow behind your conversation intelligence use case.
We will define the conversation object, agent run, webhook, and integration behavior your product needs.
Talk to founders