Skip to content
FreyaTTS-small is open source. Explore our Turkish speech model.
Book a demo

Integrations

Freya connects to your line and your systems

The agent runs on your phone line. Mid conversation it sends an HTTP request to your own systems and uses the answer in the same sentence. When the call ends we push the result to the endpoint you choose. To put the speech models inside your own product, the REST API and the WebSocket are open.

Four connection points

We build every integration through one of these four.

Phone line

The agent runs on your number. It answers inbound calls and dials outbound campaigns.

  • We connect over SIP
  • Your numbers and your carrier stay
  • Inbound calls answered
  • Outbound campaign dialling

REST API

Call the text to speech and speech to text endpoints directly.

  • TTS: /v1/audio/speech
  • STT: /v1/audio/transcriptions
  • WebSocket for live transcription
  • Bearer key auth

In flow system calls

During the call the agent sends a request to your service, waits for the answer and uses it in its reply.

  • You define the request in the flow
  • Your own auth applies
  • Response fields land in the reply
  • CRM, core banking, policy systems

Webhook

As soon as the call ends we push the result to your endpoint.

  • We fire it when the call ends
  • You set the endpoint
  • You write the result into your systems
  • Campaign calls included

What the agent reaches mid conversation

When a customer asks something, the agent does not invent the answer. It sends the request you defined in the flow, reads the data that comes back and speaks it in the same call.

1

The customer asks

A policy, a balance, an application status. The answer sits in your data.

2

The agent sends the request

The HTTP request defined in the flow goes to your service.

3

The answer enters the call

The agent uses the returned fields inside its sentence.

4

The result comes back to you

When the call ends the webhook carries the result into your systems.

What it connects to

Any system that speaks HTTP. We keep no list of ready made connectors, we build the connection over the API and webhooks.

  • CRM
  • Core banking
  • Policy and claims systems
  • Your own internal services

Integrations for either deployment option

Use the same voice agents and speech models on-prem or on cloud infrastructure in Türkiye.

Cloud hosted in Türkiye

Use Freya on cloud infrastructure located in Türkiye. Freya manages infrastructure and capacity.

On-prem

Run Freya on your own servers, including an air-gapped network when required.

The endpoints are open

Use the speech models inside your own product. Send the requests with a Bearer key.

curl -X POST "https://tts.freyavoice.ai/v1/audio/speech" \
  -H "Authorization: Bearer FREYA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"freya-tts","voice":"leyla","input":"Merhaba, ben Freya."}' \
  --output speech.wav

curl -X POST "https://stt.freyavoice.ai/v1/audio/transcriptions" \
  -H "Authorization: Bearer FREYA_API_KEY" \
  -F "file=@call.wav" \
  -F "model=freya-stt" \
  -F "language=tr"
Request API access
POST

Text to speech

https://tts.freyavoice.ai/v1/audio/speech

POST

Speech to text

https://stt.freyavoice.ai/v1/audio/transcriptions

WSS

Live transcription

wss://stt.freyavoice.ai/v1/audio/stream

A browser cannot set the Authorization header on a WebSocket. Keep the key on your own server and open the socket from there.

Let us build the connection together

Tell us which systems you need to reach and we will plan the setup.

Contact us