Every time I wanted to experiment with a new LLM, the process looked like this:
Create another account
Add another billing method
Generate another API key
Learn another pricing model
Change my code
Hope the provider doesn't go down
It wasn't hard—it was just repetitive.
As an indie developer, I wanted to spend time building products, not managing AI providers.
So I built Yantra AI.
What is Yantra AI?
Yantra AI is an AI routing gateway that lets you access multiple LLM providers through a single API key.
Instead of managing separate integrations for OpenAI, Anthropic, Gemini, DeepSeek, Mistral, Grok, Qwen, and others, you integrate once.
Your App
│
▼
Yantra AI
│
├── OpenAI
├── Anthropic
├── Gemini
├── DeepSeek
├── Mistral
└── ...
No SDK rewrites.
No juggling API keys.
No separate billing dashboards.
Why I built it
Three things kept bothering me:
client = OpenAI(
base_url="[https://api.apikeys.in/v1](https://api.apikeys.in/v1)",
api_key="YOUR_API_KEY"
) response = client.chat.completions.create(
model="claude-sonnet",
messages=[
{"role": "user", "content": "Hello!"}
]
)
That's it.
Lessons from building it
Building the routing engine was only half the challenge.
The harder part was creating a developer experience that feels invisible.
The best infrastructure is the one developers don't have to think about.
What's next?
I'm working on:
More model providers
Better routing strategies
Analytics
Team workspaces
SDKs
Streaming improvements
I'd love your feedback
If you're building with AI APIs, I'd love to know: What's the biggest pain point today?
Which providers do you use most?
What would make a routing gateway genuinely useful for you?
I'm always open to suggestions and feature requests.
Happy building! 🚀