Experiential is the open source, zero markup gateway for BYOK, self-hosted and 1000+ marketplace models. It learns from your traffic to cut costs, recommend better models, and train a specialized model you own.
We’re Kion and Silen, two AI researchers. Every time we wanted to try a new model, we had to create another account, manage another key, and track spend in another place.
We looked for an open source solution that combined model access, usage management, and a marketplace without adding another tax to every token. We couldn’t find one, so we built Experiential.
Experiential gives developers and companies one key for more than 1,000 models. You can use our marketplace, bring your own provider keys, or connect models running locally or in your cloud, all with zero token markup.
It's also more than just a gateway.
Every request shares one trace format, so the platform can catch cache misses and wasted tokens, recommend better models, identify work that should run asynchronously or in batches, and train specialized models that you own.
Since launching on Hacker News a week ago, more than 1000+ developers and 50+ companies have used Experiential to process over 10 billion tokens daily, and the repository has reached 880+ GitHub stars.
For the launch, GPT-6 Astra, DeepSeek V4 Flash, Qwen 3.8 27B, GPT-5.6 Luna, and Fable 5.1 on release are free through the end of the week 🎁
Experiential Labs
Hey Product Hunt 👋
We’re Kion and Silen, two AI researchers. Every time we wanted to try a new model, we had to create another account, manage another key, and track spend in another place.
We looked for an open source solution that combined model access, usage management, and a marketplace without adding another tax to every token. We couldn’t find one, so we built Experiential.
Experiential gives developers and companies one key for more than 1,000 models. You can use our marketplace, bring your own provider keys, or connect models running locally or in your cloud, all with zero token markup.
It's also more than just a gateway.
Every request shares one trace format, so the platform can catch cache misses and wasted tokens, recommend better models, identify work that should run asynchronously or in batches, and train specialized models that you own.
Since launching on Hacker News a week ago, more than 1000+ developers and 50+ companies have used Experiential to process over 10 billion tokens daily, and the repository has reached 880+ GitHub stars.
For the launch, GPT-6 Astra, DeepSeek V4 Flash, Qwen 3.8 27B, GPT-5.6 Luna, and Fable 5.1 on release are free through the end of the week 🎁
We’d love your feedback - experientiallabs.ai
And a star - github.com/experientiallabs/experiential