Apiarium
One API key for GPT, Claude, image gen, TTS & transcription
Quick update since my launch post here a couple weeks ago. Been heads-down shipping based on what I've learned so far:
New AI provider: Added Google Gemini alongside GPT-4o and Claude, including Gemini's new image generation model (Nano Banana), so now you can generate images with either OpenAI or Google models through the same endpoint.
Multi-currency support: Pricing now auto-detects and displays in 10 currencies (USD, EUR, GBP, CAD, INR, AUD, BRL, MXN, JPY, SGD) instead of forcing everyone into USD.
Smarter credit system: Reworked how credits work for subscribers, pay-as-you-go top-ups are now priced per plan, with better rates the higher your tier, so upgrading actually saves you money on extra usage instead of just adding a fixed fee.
Cleaner error handling: Normalized error responses across all 4 providers, so instead of parsing OpenAI's error format one way and Anthropic's another, you get one consistent error structure regardless of which model you called.
Still solo, still building based on feedback, if you tried Apiarium after the launch post, would love to know what's missing or what would make you switch from your current setup.
Hey IH 👋
I got tired of managing 4 different API accounts every time I wanted to add AI to a project. OpenAI for text and images, Anthropic for Claude, separate billing for each...
So I built Apiarium. One API key, all the models, credit-based pricing so you always know what you're spending.
It supports:
- GPT-4o-mini, GPT-4o, Claude Haiku, Claude Sonnet
- Image generation
- Text to speech (6 voices)
- Audio transcription (Whisper)
and more updates coming soon
Free tier to start, no credit card required.
Would love feedback from other devs, especially on website, pricing and what models/features you'd want next.
-
2
The interesting part to me isn't having one API key.
It's becoming the layer developers choose to build against.
Once someone integrates your API, switching costs start working in your favor.
I'd be curious whether that's the outcome you're optimizing for, or whether simpler billing is the bigger pull today.
-
1
Honest answer: right now simpler billing is the bigger pull. That's the pain I felt and built around.
But you're pointing at something real. The switching cost dynamic is what makes this interesting long-term, once someone builds their image pipeline or transcription workflow against Apiarium, they're not re-integrating four separate providers to save a few dollars.
The question I'm still figuring out is whether I need to optimize for that explicitly or whether it just happens as a byproduct of being reliable and keeping the API stable. Curious what you've seen work.
-
Comment