This is the new api endpoint that unifies some available and some custom back end services.
The point is with so many emerging ai models and technologies developers need a curated simple solution.
If every body is hacking together their own solutions time and energy is being wasted on the backend.
What if there was an api available that can do the advance stuff out of the box.
All the developer would need to do is build a frontend for their client and deliver world class ai inference and insight into their domain of choice.
As a phase one trial we will be incorporating a few technologies
- LLM: large language models
- OpenAI chat GPT 3.5
- These are trained on large amounts of data and can provide deep contextual insights given enough context
- Embedding: semantic text embeddings
- Instruct Embedding - allows general use specific semantic text vector embeddings
- These allow us to encode text information in a vector for fast similarity search
- Vector Store
- QDrant: fast rust based vector db and similarity search engine
- In order to manager and search our vector data we need a place to put it
We host all of these severs in a scalable monolithic backend.
We then simplify them into a globally chaced and fast edge front end for hosting your chats.
Finally, we add a user friendly billing system and auth layer on top of this for seamlessly building your apps.
This is more then just an ai api. this is an ai app platform all in one. You have the freedom to offer a world class service for your company
Its simple stick