Skip to content

Commit 4929fd5

Browse files
author
Abhishek
committed
Add Telnyx integration example with AI services support
This commit adds documentation and configuration examples for using Telnyx AI services with EchoKit Server. Why Telnyx: - OpenAI-compatible API enables drop-in integration - 53+ AI models including GPT-4, Claude, and Llama - Global edge network for low-latency inference - Unified API key for ASR, TTS, and LLM services - Pay-per-use pricing with transparent costs Changes: - examples/telnyx/config.toml: Ready-to-use configuration - examples/telnyx/README.md: Complete integration guide
1 parent fe01bbd commit 4929fd5

2 files changed

Lines changed: 205 additions & 0 deletions

File tree

‎examples/telnyx/README.md‎

Lines changed: 141 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,141 @@
1+
# Telnyx Integration for EchoKit Server
2+
3+
This guide explains how to configure EchoKit Server to use Telnyx AI services.
4+
5+
## Why Telnyx?
6+
7+
Telnyx provides a comprehensive AI platform that integrates seamlessly with EchoKit:
8+
9+
- **OpenAI-Compatible API**: All endpoints follow OpenAI specifications, enabling drop-in compatibility with EchoKit's architecture
10+
- **53+ AI Models**: Access to GPT-4, Claude, Llama, Mistral, and many open-source models through a single API
11+
- **Global Edge Network**: Low-latency inference from data centers worldwide
12+
- **Unified Billing**: Single API key for ASR, TTS, and LLM services
13+
- **Competitive Pricing**: Pay-per-use with transparent, per-token pricing
14+
15+
## Prerequisites
16+
17+
1. A Telnyx account ([sign up here](https://telnyx.com/sign-up))
18+
2. An API key from the [Telnyx Portal](https://portal.telnyx.com)
19+
20+
## Quick Start
21+
22+
### 1. Set Your API Key
23+
24+
```bash
25+
export TELNYX_API_KEY="your-api-key-here"
26+
```
27+
28+
### 2. Use the Example Configuration
29+
30+
```bash
31+
cp examples/telnyx/config.toml config.toml
32+
```
33+
34+
### 3. Build and Run
35+
36+
```bash
37+
cargo build --release
38+
./target/release/echokit_server
39+
```
40+
41+
## Configuration Reference
42+
43+
### ASR (Speech Recognition)
44+
45+
```toml
46+
[asr]
47+
platform = "openai"
48+
url = "https://api.telnyx.com/v2/ai/transcriptions"
49+
api_key = "${TELNYX_API_KEY}"
50+
model = "whisper-1"
51+
lang = "en"
52+
```
53+
54+
Available models:
55+
- `whisper-1` - OpenAI Whisper (recommended)
56+
57+
### TTS (Text-to-Speech)
58+
59+
```toml
60+
[tts]
61+
platform = "openai"
62+
url = "https://api.telnyx.com/v2/ai/speech"
63+
model = "tts-1"
64+
api_key = "${TELNYX_API_KEY}"
65+
voice = "alloy"
66+
```
67+
68+
Available voices:
69+
- `alloy`, `echo`, `fable`, `onyx`, `nova`, `shimmer`
70+
71+
Available models:
72+
- `tts-1` - Optimized for speed
73+
- `tts-1-hd` - Higher quality audio
74+
75+
### LLM (Language Model)
76+
77+
```toml
78+
[llm]
79+
platform = "openai_chat"
80+
url = "https://api.telnyx.com/v2/ai/chat/completions"
81+
api_key = "${TELNYX_API_KEY}"
82+
model = "gpt-4o-mini"
83+
history = 5
84+
```
85+
86+
Popular model options:
87+
- `gpt-4o` - Latest GPT-4 optimized
88+
- `gpt-4o-mini` - Fast, cost-effective
89+
- `claude-3-5-sonnet` - Anthropic Claude
90+
- `llama-3.1-70b-instruct` - Open-source alternative
91+
- `llama-3.1-8b-instruct` - Lightweight, fast inference
92+
93+
See the [Telnyx AI documentation](https://developers.telnyx.com/docs/ai/introduction) for the complete model list.
94+
95+
## Using Telnyx LiteLLM Proxy
96+
97+
For advanced use cases, Telnyx offers a LiteLLM proxy that provides:
98+
99+
- Automatic fallback between models
100+
- Load balancing across providers
101+
- Unified rate limiting
102+
- Custom model aliases
103+
104+
Configure the proxy URL in your `config.toml`:
105+
106+
```toml
107+
[llm]
108+
platform = "openai_chat"
109+
url = "https://api.telnyx.com/v2/ai/chat/completions"
110+
# Use any supported model
111+
model = "claude-3-5-sonnet"
112+
```
113+
114+
## Troubleshooting
115+
116+
### Authentication Errors
117+
118+
Verify your API key is set correctly:
119+
120+
```bash
121+
echo $TELNYX_API_KEY
122+
```
123+
124+
### Model Not Found
125+
126+
Check available models in your Telnyx Portal or consult the [API documentation](https://developers.telnyx.com/docs/ai/introduction).
127+
128+
### High Latency
129+
130+
Telnyx routes requests to the nearest edge location automatically. If you experience latency issues, verify your network connection or contact Telnyx support.
131+
132+
## Additional Resources
133+
134+
- [Telnyx AI Documentation](https://developers.telnyx.com/docs/ai/introduction)
135+
- [Telnyx API Reference](https://developers.telnyx.com/docs/api/v2/overview)
136+
- [EchoKit Documentation](https://echokit.dev/docs/quick-start/)
137+
- [Telnyx Discord Community](https://discord.gg/telnyx)
138+
139+
## License
140+
141+
This integration example is provided under the same license as EchoKit Server.

‎examples/telnyx/config.toml‎

Lines changed: 64 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,64 @@
1+
# EchoKit Server Configuration with Telnyx
2+
#
3+
# This example demonstrates using Telnyx AI services with EchoKit Server.
4+
# Telnyx offers 53 AI models via an OpenAI-compatible inference API,
5+
# making it a natural fit for EchoKit's OpenAI-spec architecture.
6+
#
7+
# Key benefits:
8+
# - Single API key for LLM, ASR, and TTS
9+
# - Global edge network for low-latency inference
10+
# - 53+ AI models including GPT, Claude, Llama, and open-source options
11+
# - Pay-per-use pricing with no minimum commitments
12+
#
13+
# Setup:
14+
# 1. Create a Telnyx account at https://telnyx.com
15+
# 2. Generate an API key from the Portal
16+
# 3. Set your TELNYX_API_KEY environment variable
17+
#
18+
# API Documentation: https://developers.telnyx.com
19+
20+
addr = "0.0.0.0:8080"
21+
hello_wav = "hello.wav"
22+
23+
[asr]
24+
platform = "openai"
25+
url = "https://api.telnyx.com/v2/ai/transcriptions"
26+
api_key = "${TELNYX_API_KEY}"
27+
model = "whisper-1"
28+
lang = "en"
29+
30+
[tts]
31+
platform = "openai"
32+
url = "https://api.telnyx.com/v2/ai/speech"
33+
model = "tts-1"
34+
api_key = "${TELNYX_API_KEY}"
35+
voice = "alloy"
36+
37+
[llm]
38+
platform = "openai_chat"
39+
url = "https://api.telnyx.com/v2/ai/chat/completions"
40+
api_key = "${TELNYX_API_KEY}"
41+
model = "gpt-4o-mini"
42+
history = 5
43+
44+
[[llm.sys_prompts]]
45+
role = "system"
46+
content = """
47+
You are a helpful assistant. Answer truthfully and concisely. Always answer in English.
48+
49+
- NEVER use bullet points
50+
- NEVER use tables
51+
- Answer in complete English sentences as if you are in a conversation.
52+
53+
"""
54+
55+
# Alternative: Use Telnyx's LiteLLM proxy for unified access to 53+ models
56+
# Uncomment the section below and comment out the [llm] section above
57+
#
58+
# [llm]
59+
# platform = "openai_chat"
60+
# url = "https://api.telnyx.com/v2/ai/chat/completions"
61+
# api_key = "${TELNYX_API_KEY}"
62+
# # Available models include: gpt-4o, claude-3-5-sonnet, llama-3.1-70b, and more
63+
# model = "llama-3.1-70b-instruct"
64+
# history = 5

0 commit comments

Comments
 (0)