Prefer the interface? Configure Voice in Studio. Both methods update the same Voice configuration.
Before you begin
- A server-side Voiceflow API key with project version-read and project-update permissions.
- The Environment’s concrete version ID; this route does not resolve a display name or alias. Find it through Environment APIs.
- Existing Voice settings on that Environment,
curlandjq.
This guide uses the exact v1alpha1 Voice settings contract. Do not substitute the stable Environment resource route.
Update the configuration
Read and retain the current settings
The response wraps configuration in voice. Keep this file to compare and restore your change. A 404 can mean that the Environment has no Voice settings, not just an incorrect ID.
Build a focused patch
This example sets a 30-minute maximum call duration. These are illustrative values, not defaults or recommended limits. PATCH merges top-level fields, but replaces supplied nested objects except keypadInput; merge existing stt, tts or maxDuration before changing a single nested field.
Read back and test
Repeat the GET into a new file and compare it with voice-before.json. A successful PATCH returns 204 with no JSON body. Call the draft over your actual phone or custom client and test the changed behavior before publishing.
Verify the result
Confirm the returned settings contain your patch and retain unrelated provider options. Test that the duration limit and its End message apply to the intended call. A saved field does not prove that your intended Environment answered the test call.
Restore or publish
PATCH only the changed fields from voice-before.json to restore them. Resolve 400 schema errors or 403 permission errors before retrying. Keep keys on your server. Use Release a voice integration for publishing and routing.