> ## Documentation Index
> Fetch the complete documentation index at: https://docs.eigi.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Supported Providers

> Explore the wide range of AI providers eigi.ai supports for language models, voice synthesis, speech recognition, and video avatars

## Overview

eigi.ai integrates with leading AI providers to deliver the best possible experience for your voice agents. Mix and match providers to create the perfect combination for your use case.

<Info>
  All provider configurations are managed through the no-code dashboard—simply
  select your preferred options and deploy.
</Info>

***

## Language Models (LLM)

Power your agent's intelligence with industry-leading language models:

<CardGroup cols={2}>
  <Card title="OpenAI" icon="circle-o">
    **GPT-4o, GPT-4o Mini** Industry-leading models with exceptional reasoning,
    instruction following, and conversation capabilities.
  </Card>

  <Card title="Anthropic" icon="circle-a">
    **Claude 3.5 Sonnet, Claude 3 Opus** Known for nuanced understanding,
    safety, and handling complex instructions with precision.
  </Card>

  <Card title="Google" icon="circle-g">
    **Gemini 2.0, Gemini 1.5 Pro** Multimodal capabilities with strong
    performance across diverse conversation types.
  </Card>

  <Card title="Groq" icon="bolt">
    **LLaMA 3, Mixtral** Ultra-fast inference for latency-critical applications
    requiring quick responses.
  </Card>
</CardGroup>

### Model Selection Guide

| Use Case             | Recommended Model         | Why                                    |
| -------------------- | ------------------------- | -------------------------------------- |
| **Complex Support**  | GPT-4o, Claude 3.5 Sonnet | Superior reasoning for troubleshooting |
| **High Volume**      | GPT-4o Mini, Groq         | Balance of speed and cost efficiency   |
| **Sensitive Topics** | Claude 3.5 Sonnet         | Nuanced, careful responses             |
| **Fast Responses**   | Groq (LLaMA 3)            | Lowest latency for real-time feel      |

***

## Voice Synthesis (TTS)

Transform your agent's responses into natural, expressive speech:

<CardGroup cols={2}>
  <Card title="ElevenLabs" icon="waveform">
    **Premium Voice Quality** Ultra-realistic voices with emotional range,
    natural pacing, and voice cloning capabilities. - 30+ languages - Custom
    voice cloning - Expressive speech styles
  </Card>

  <Card title="Cartesia" icon="wave-pulse">
    **Low-Latency Synthesis** Optimized for real-time conversations with minimal
    delay and natural sound. - Sub-100ms latency - Streaming synthesis -
    Multiple voice options
  </Card>

  <Card title="Sarvam" icon="globe">
    **Indian Languages** Specialized support for Indian languages with authentic
    accents and regional variations. - Hindi, Tamil, Telugu, and more - Native
    speaker quality - Regional dialect support
  </Card>

  <Card title="OpenAI" icon="circle-o">
    **Integrated Solution** Reliable voice synthesis integrated with OpenAI's
    ecosystem. - Consistent quality - Multiple voice personas - Easy
    configuration
  </Card>
</CardGroup>

### Voice Provider Comparison

| Provider       | Best For                              | Latency | Voice Cloning |
| -------------- | ------------------------------------- | ------- | ------------- |
| **ElevenLabs** | Premium quality, emotional expression | Medium  | ✅ Yes         |
| **Cartesia**   | Real-time, low latency conversations  | Low     | ❌ No          |
| **Sarvam**     | Indian language support               | Medium  | ❌ No          |
| **OpenAI**     | Simple, reliable deployment           | Medium  | ❌ No          |

***

## Speech Recognition (STT)

Convert user speech to text with high accuracy:

<CardGroup cols={2}>
  <Card title="Deepgram" icon="ear-listen">
    **Industry-Leading Accuracy** Real-time speech recognition with exceptional
    accuracy across accents and environments. - Nova-2 model for best accuracy -
    Real-time streaming - Multi-language support - Background noise handling
  </Card>

  <Card title="Azure Speech" icon="microsoft">
    **Enterprise Grade** Microsoft's speech recognition with enterprise features
    and compliance. - 100+ languages - Custom speech models - Enterprise
    security
  </Card>
</CardGroup>

### STT Features

| Feature                 | Deepgram   | Azure Speech |
| ----------------------- | ---------- | ------------ |
| **Real-time Streaming** | ✅          | ✅            |
| **Noise Cancellation**  | ✅ Advanced | ✅            |
| **Custom Vocabulary**   | ✅          | ✅            |
| **Speaker Diarization** | ✅          | ✅            |
| **Punctuation**         | ✅ Auto     | ✅ Auto       |

***

## Video Avatars

Create visual representations for immersive video conversations:

<CardGroup cols={2}>
  <Card title="HeyGen" icon="user-tie">
    **Professional Avatars** High-quality digital humans for professional and
    corporate use cases. - 100+ stock avatars - Custom avatar creation -
    Professional appearance - Multiple languages
  </Card>

  <Card title="Tavus" icon="video">
    **Personalized Video** Create personalized video experiences with
    AI-generated presenters. - Custom avatar training - Brand consistency -
    Scalable video generation
  </Card>
</CardGroup>

### Avatar Provider Comparison

| Provider   | Stock Avatars | Custom Avatars | Best For                |
| ---------- | ------------- | -------------- | ----------------------- |
| **HeyGen** | 100+          | ✅ Yes          | Professional, corporate |
| **Tavus**  | Limited       | ✅ Yes          | Personalized, branded   |

***

## Telephony

Connect your agents to phone networks:

<CardGroup cols={2}>
  <Card title="Plivo" icon="phone">
    **Global Coverage** Enterprise-grade telephony with global phone number
    availability. - 190+ countries - Inbound & outbound - SIP trunking
  </Card>

  <Card title="Twilio" icon="phone-volume">
    **Developer Friendly** Flexible telephony platform with extensive API
    capabilities. - Programmable voice - Global reach - Carrier network
  </Card>
</CardGroup>

***

## Choosing the Right Stack

### For Customer Support

```
LLM: GPT-4o or Claude 3.5 Sonnet (complex reasoning)
TTS: ElevenLabs (empathetic, natural voice)
STT: Deepgram Nova-2 (accurate transcription)
```

### For High-Volume Sales

```
LLM: GPT-4o Mini or Groq (cost-effective, fast)
TTS: Cartesia (low latency)
STT: Deepgram (reliable streaming)
```

### For Indian Markets

```
LLM: GPT-4o (multilingual capability)
TTS: Sarvam (native Indian language voices)
STT: Deepgram (Hindi/regional language support)
```

### For Video Experiences

```
LLM: Claude 3.5 Sonnet (nuanced conversation)
TTS: ElevenLabs (expressive synthesis)
STT: Deepgram Nova-2 (accurate capture)
Avatar: HeyGen or Tavus (professional appearance)
```

***

## Provider Availability

<Tip>
  All providers are available through the eigi.ai dashboard. Simply configure
  your agent and select your preferred providers—no API keys or technical setup
  required.
</Tip>

<Note>
  Provider availability and features may vary based on your plan. Contact
  support for enterprise options and custom provider integrations.
</Note>

<Card title="Need a Different Provider?" icon="envelope" href="mailto:buddy@eigi.ai">
  We're constantly adding new providers. Let us know what you need!
</Card>