Stringing together a range of AI models to do anything useful is hard. There isn't an accessible and unified platform that pulls together all the disparate technologies in one place.

This means that only developers can easily access bleeding edge tech. The tools they use are not user-friendly and are often only made for single use cases.

In the future, though, we believe software will move beyond AI-first applications that address those single use cases. Instead, the best software will empower users to build multimodal, tailored workflows that combine the best of machine learning, natural language processing, computer vision, and artificial intelligence.

That’s what Odyssey is for. It’s not a single-use, AI wrapper but an all-purpose piece of software that lets any creator leverage AI to create great art, get work done, or automate repetitive tasks.

How did we get here?

Odyssey is a team of three: one engineer, one designer, and one product manager. We met working in big tech where we were all passionate about taking complex technology and making it extremely approachable for a non-technical audience.

We started working together when generative image models were first becoming popular. We built an iOS app called Creator that was focused on making image generation seamless and simple for a non-technical audience. 

When we started building, there were two AI image generation apps on Apple’s app store. They had clunky UIs, and required complex prompts. 

By the time we launched (about 3-4 months later), there were dozens, if not hundreds. We built an app we were proud of, but ostensibly missed the hype cycle.

When we started planning what to work on next, we aligned on three key principles:

1. The power is in the workflow - we are not building an image generator or a chatbot creator. While you can use Odyssey to do those things, there are other apps that make that possible as well. Instead, we intend Odyssey to have a full spectrum of AI and ML models, APIs, image filters, text transformers, and utility actions that can all be stitched together to achieve a desirable result
   
   

2. Models should run locally whenever possible - If you pay to run an AI model in the cloud, you are either paying a margin on top of the true cost to run a model or being subsidized by VC money. Long-term, this is not sustainable. While it’s harder to run models locally, we believe it’s the future. That means with Odyssey you’re paying for a beautifully designed, thoughtfully built piece of software from a team that loves and cares about what they’re doing - not the cost of power hungry GPUs
   
   

3. Build for the long-term - the current state of AI means there are new trends every week. While it’s tempting to go after the latest, the nature of trends is that they end. We want to build something long lasting that can survive the AI hype cycles.

We are not there yet.

Our name (Odyssey) is a nod to the journey that is ongoing and, if we build this software correctly, we’ll continue to be on for years. (We also think the nodes look like islands in the Mediterranean sea - but that’s more abstract.)

While Odyssey is not in a complete state, we believe that creating a piece of software that is principled, considered, and opinionated is a worthwhile journey.

We hope you'll join us.