Odyssey roadmap

Our goal is to continually add the latest AI models and workflows to Odyssey as the technology continues to advance. Our roadmap page will be the best reflection of what we're working on adding to the app.

Suggest a feature

Recently added

Features that were released with the latest version of Odyssey


Workflows

Save, style, and share workflows. Workflows are one of the last pieces of core functionality before we have a full, working application.

LLM improvements

We are making minor bug fixes and building a more consistent experience for Llama2 and ChatGPT.

Undo/redo

Undo/redo is a great example of how a GUI app is a big UX win over command-line utilities or trying to move assets manually between different, single-purpose apps. This is a huge doubling down on the strength of Odyssey: it encourages the user to experiment and string ideas together.

Model upload

Support a wider range of fine tuned models - particularly more popular models on platforms like CivitAI.

Local SDXL

Run SDXL locally.

In progress

Features that are actively in development

Billing

Odyssey will offer a 14-day free trial and then cost $11.99/month when paid monthly and $9.99/month when paid annually.

Local SDXL ControlNet

ControlNet support for SDXL is still in progress.

Fine tuning

We are adding the ability to fine tune SDXL via API.

Improved upscaling

Add new super resolution models to Odyssey. We are adding an update version of ESRGAN, GFPGAN, and potentially a ControlNet Tile upscaling workflow.

Coming soon

We're listening and we're working on it. Expect these updates in the coming weeks or months.

Mistral 7B support

We are adding the Mistral LLM to Odyssey.

Transcription, translation, and file conversion

Additional nodes for audio transcription, multi-language translation, and converting files.

Multi-modal input

Input PDFs, CSVs, and unstructured data into our LLM nodes. Improved support of larger files as context.

Prompt book

Reference, load, and pre-fill prompts for both Stable Diffusion and LLMs directly from Odyssey.

Video support

We are building a video workflow for Odyssey.

API node

An unstructured API node that allows advanced users to create workflows that support a wide range of developer APIs.

Help wanted

We want to add these features to Odyssey but a technical foundation needs to happen first.

Latent consistency model

We want Odyssey to run a latent consistency model of SDXL locally. While it's technically possible to convert a model to CoreML and do this now, it would take us a significant amount of time and distract from our roadmap. If you're interested in converting this model to CoreML for us, get in touch!

Segment Anything

Similarly we'd love to add Meta's Segment Anything model to Odyssey but we'd need the model converted to CoreML first. If you're interested in converting this model to CoreML for us, get in touch!

Audio support

We love Stable Audio and we're hoping for the ability to run the model locally. If that's not possible, we'll jump on an API. If you know of any equivalent audio models that can run locally on a Mac, get in touch.


In progress

Features that are actively in development


Workflows

Save, style, and share workflows. Workflows are one of the last pieces of core functionality before we have a full, working application.

Billing

Odyssey will offer a 14-day free trial and then cost $11.99/month when paid monthly and $9.99/month when paid annually.

LLM improvements

We are making minor bug fixes and building a more consistent experience for Llama2 and ChatGPT.

Undo/redo

Undo/redo is a great example of how a GUI app is a big UX win over command-line utilities or trying to move assets manually between different, single-purpose apps. This is a huge doubling down on the strength of Odyssey: it encourages the user to experiment and string ideas together.