Every seat you sell is quietly building a structured data layer across your base. The API to build on it is already in the seat, at every tier, at no extra charge. This page is what that unlocks.
This is what your tech lead wants to know.
The API is live and callable today: JSON over HTTPS, versioned under /api/v1. Authenticate once, then read processed conversations, their analysis, and the alerts raised on them. The products in the seat are consumers of this same API. Anything they show, your systems can retrieve.
| Read surface | Returns |
|---|---|
GET /calls | Filterable conversation list, with summary KPI counts |
GET /calls/{id} | Full detail: metadata, summary, embedded tasks, the analysis objects, the vCon |
GET /alerts | Evidence-linked moments flagged within minutes of hang-up, with KPIs and top triggers |
Bearer tokens, server-side tenant scoping, cursor pagination, a consistent error envelope. New fields are additive within a version.
| Events out | What fires |
|---|---|
After Call webhook | Per completed conversation: summary, handling user, metadata, sentiment |
First Alert webhook | When an alert condition matches: a keyword or phrase, a sentiment threshold, or a topic by name, with the matched criteria and the evidence attached |
POST /cdrs | Submit conversations for processing directly. Asynchronous |
Ingestion access, signing secrets, and rate limits are provisioned per integrator. A dedicated Developer Studio is on the roadmap.
The pattern is consistent: you bring the customer-facing application and the experience. The layer brings the understanding of what was said and what it means, already tuned to the industry, so your app is industry-aware on day one instead of after a year of its own tuning.
An agent reading Tresic classification treats a caller describing symptoms differently from a routine scheduling request, and sizes a home-services lead by job scope and revenue potential. Your app decides what to do next. The layer tells it what the conversation is.
Follow-ups generated from the conversation object itself, referencing what was actually discussed, in the right terminology for the industry.
Classify intent, urgency, and risk from the first sentence, so your routing logic can send an emergency to dispatch instead of the scheduling queue. No static menu trees.
Automation built on what a conversation means in context: a routine inquiry versus a bad-faith claim, a filter change versus a system replacement.
The seat pays its way from day one. The layer underneath it is where the next products come from.
The seat, attached to the phones you already sell. Valuable the day it turns on, and the data layer starts populating across your base with every conversation.
Voice, messaging, workflow, and deeper reporting built on the API, on a layer that already exists across your base, at preferred partner rates.