At Replit, we continue to ship features with a unique north star in mind: providing the shortest path from idea to software running in production. Last week’s Autoscaling and Static Deployments announcement was another key step in building an end-to-end platform that supports scalable, production-ready projects. Today, we announce Replit ModelFarm, the fastest and safest way to build Generative AI applications.
You can now build Gen AI apps without the hassle of API key management – Replit ModelFarm natively supports secure access to 3rd-party Gen AI providers. Developers will be able to focus only on building, without having to worry about their API keys leaking. In the future, we will also provide an integrated billing system with support for quota limits, alarms, and more. And if you want to use any additional APIs, our Secrets feature makes it safe and straightforward.
Writing your first Gen AI app is as easy as:
from replit.ai.modelfarm import CompletionModel
model = CompletionModel("text-bison")
prompts = ["What is the best way to learn AI?"]
completion = model.complete(prompts, max_output_tokens=50, temperature=0.2)
The speed of adoption of Gen AI has reached a scale never witnessed before in the tech world. In our recent State of AI Development post, we report 80% quarter-on-quarter growth in the number of Gen AI projects created on Replit. We don’t expect this growth to slow down anytime soon – if anything, the energy that AI builders have been pouring into their projects never ceases to amaze us. Now, we want every software creator to turn into an AI software creator.
Generative AI has sparked the imagination of millions already. We can’t wait to see what you will build with Replit ModelFarm!