API Reference

Resources

API Reference & SDK

Complete REST API & WebSocket documentation for Vocred

Quickstart

All API requests require a JWT token in the Authorization header. Include your API key as the Bearer token.

curl -X GET "https://api.vocred.ai/v1/agents" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json"

Authentication API

4 endpoints

WebSocket Events

Events sent between the client and server during a voice session over wss://api.vocred.ai/ws/voice

Direction	Event	Description
server → client	ready	Voice session is ready. Server sends session config.
server → client	transcript	Real-time speech-to-text transcript from the user.
server → client	response_text	Full AI response text when generation completes.
server → client	response_text_chunk	Streaming chunk of AI response text.
server → client	audio	Full audio response (base64-encoded).
server → client	audio_stream	Binary audio stream chunk (raw PCM16).
server → client	audio_end	Signals end of audio response.
server → client	session_ended	Voice session has ended with reason.
server → client	interrupt	User interruption detected; server stops playback.
server → client	clear_audio	Clear audio playback buffer.
server → client	error	Server-side error occurred.
client → server	audio	User microphone audio (base64-encoded PCM16).
client → server	audio_binary	User microphone audio as raw binary PCM16.
client → server	interrupt	Client-side interrupt signal (user barge-in).
client → server	end_call	Request to end the voice session.

Models & Endpoints

LLM Providers

GroqFree tier, ultra-fast

llama-3.1-8b-instant, llama-3.3-70b-versatile, mixtral-8x7b-32768

OpenAIPaid, industry standard

gpt-4o, gpt-4o-mini, gpt-4-turbo

GeminiFree tier available

gemini-1.5-flash, gemini-1.5-pro

AnthropicSafety-focused

claude-3-haiku, claude-3-sonnet

CerebrasUltra-fast inference

llama-3.1-8b, llama-3.3-70b

Voice & Speech

Edge TTSFree, unlimited

Jenny, Guy, Neerja, Swara, 100+ voices

ElevenLabsHD, 10k chars/mo free

Eleven Turbo v2, Eleven Multilingual v2

CartesiaLow latency

Sonic (ultra-fast)

Azure TTSEnterprise quality

50+ neural voices

Deepgram$200 free credits

Nova-2, Whisper

Error Codes

Code	Error	Description
400	Bad Request	Invalid request parameters or body
401	Unauthorized	Missing or invalid JWT token
403	Forbidden	Insufficient permissions
404	Not Found	Resource does not exist
409	Conflict	Resource already exists (e.g., duplicate name)
422	Unprocessable Entity	Validation error in request body
429	Too Many Requests	Rate limit exceeded. Retry after 60s.
500	Internal Server Error	Something went wrong on our end
502	Bad Gateway	Upstream provider (LLM/STT/TTS) unavailable
503	Service Unavailable	Service temporarily down for maintenance

Client SDK (npm)

Install the official Vocred SDK for programmatic access to all API endpoints.

npm install vocred-sdk

import { Vocred } from 'vocred-sdk'

const vocred = new Vocred({ apiKey: 'YOUR_API_KEY' })

// List agents
const agents = await vocred.agents.list()

// Create agent
const agent = await vocred.agents.create({
  name: 'Support Bot',
  systemPrompt: 'You are a helpful agent...',
  voiceProvider: 'edge-tts',
  llmProvider: 'groq',
})

// Start a voice call
const call = await vocred.calls.create({
  to: '+919999999999',
  agentId: agent.id,
})