#Getting started
This installation of the platform runs at:
| What | URL |
|---|---|
| API base URL | https://tvh-voice-api.gigaleads.nl |
| Dashboard (API keys, calls, settings) | https://tvh-voice.gigaleads.nl |
| Test console (a working example integration) | https://tvh-voice-demo.gigaleads.nl |
| OpenAPI 3.1 document | https://tvh-voice-api.gigaleads.nl/openapi.yaml |
| All documentation in one file (for AI coding agents) | https://tvh-voice-api.gigaleads.nl/llms-full.txt |
#1. Get an API key
An admin creates a key in the dashboard. Give it only the scopes your integration needs (see API conventions). The key belongs in your server-side configuration; browser code never sees it.
export VOICE_API_URL=https://tvh-voice-api.gigaleads.nl
export VOICE_API_KEY=vk_...
curl -sS "$VOICE_API_URL/v1/agents" -H "Authorization: Bearer $VOICE_API_KEY"
#2. Build the integration
Follow the integration guide step by step: agents and the browser softphone, click-to-call, webhooks, inbound routing and call logs. Every step links to a recipe with tested code.
#3. Compare with the test console
The test console is a small integration that uses this API exactly like your backend would: one API key, a browser softphone, click-to-call, signed webhooks and recording playback. Its live log shows every request it sends and every webhook it receives — useful to compare with your own traffic.
#Notes for this installation
- Requests must send a
User-Agentheader. Most HTTP clients do; Python'surllibdefault is blocked by the network edge, so set your own (for exampleUser-Agent: my-crm/1.0). - The TypeScript SDK (
@voice/sdk) and the softphone package (@voice/softphone) are not published to npm yet. Use plain HTTPS plus any Standard Webhooks library, and load the browser softphone from /downloads/softphone.js (globalVoiceSoftphone, see Browser softphone). - Ask a question with the Ask AI button: it answers from this documentation and links the pages it used.