Skip to content
Koldan logo Koldan Documentation

One Platform,
The Perfect Solution.

Turn speech into text, insight, and action with Koldan, Dixilang’s voice AI platform. Build with transcription, live streaming, summaries, translation, and speaker intelligence.

Get started Explore the APIs

Built for developers.

team-meeting.wavExample workflow
Any file. Any format. Transcription made simple.

TRANSCRIPT English

00:04
Speaker 1

Have you sent the document to the client yet?

00:09
Speaker 2

Not yet. I’m reviewing it one last time.

00:15
Speaker 1

Okay, let me know once you’ve sent it.

How it works

THE PLATFORM

More than words on a page.

From the first spoken word to the next action.
Find the building blocks for your workflow.

Transcribe audio

Turn recordings into text with timestamps, punctuation, and speaker diarization.

Transcription API

Stream in real time

Bring live speech-to-text into your application with WebSocket streaming.

Streaming guide

Find the key takeaways

Create summaries, decisions, and action items with configurable prompt templates.

Summaries API

Work across languages

Translate transcripts into your target language and connect multilingual workflows.

Translations API

Build speaker intelligence

Enroll speakers, manage voice prints, and verify speaker identity from audio.

Speaker services

Keep control of access

Connect your identity provider and manage API keys, permissions, and usage policies.

Platform guides

Language, model, and feature availability depend on your installation. Explore available models

FROM IDEA TO INTEGRATION

Your audio.
Your stack. Your API.

Use REST for file workflows and WebSockets for live audio. Start with a request, then choose the SDK that fits your application.

  1. 01Set up accessUse an API key or your identity provider.
  2. 02Send your first recordingUpload audio and create a transcription job.
  3. 03Build with the resultWait for completion and retrieve the transcript.
Upload & transcribe cURL · Bash
# Set your server URL and API key first.
curl --fail-with-body \
  "${KOLDAN_BASE_URL}/api/v1/speech-services/transcriptions/upload" \
  -H "X-API-Key: ${KOLDAN_API_KEY}" \
  -F "file=@meeting.wav" \
  -F "name=team-meeting" \
  -F 'transcription={
    "model": "general"
  }'

This request creates a job. Configure webhooks to receive completion notifications, or poll until COMPLETED, FAILED, or CANCELLED. Use a model available on your server.

Follow the complete quickstart

BUILT FOR YOUR ENVIRONMENT

Powerful AI. On your terms.

Explore operations

Choose where it runs

Dixilang supports cloud, on-premises, and air-gapped environments. Plan a deployment that fits your organization.

Deployment requirements

Connect your controls

Integrate identity, configure TLS, and define access policies for the people and services using Koldan.

Identity and access

Operate with visibility

Follow guides for health monitoring, capacity planning, backups, and upgrades throughout your installation’s lifecycle.

Health and monitoring

THE PEOPLE BEHIND KOLDAN

Speech expertise.
Built into every layer.

Dixilang is an Israeli company developing speech and language technology since 2012. Its work spans Hebrew, Arabic, and English, helping organizations turn conversations into useful knowledge.

From healthcare and government to legal teams and contact centers, Dixilang builds for the language, terminology, and deployment needs of each organization.

Meet Dixilang

Make something worth listening to.

Your first integration starts here.

Get started with Koldan