Argil's platform currently relies on 2 base models: OpenAI's API for GPT4/GPT3.5, that is SOC2 and RGPD compliant. and does not use input data to train their model.
On the image front, we rely on an open source model of Stable Diffusion (from Stability.ai), with cloud servers in Europe - your data is safe here.
In the future, we will integrate other LLMs (including open source ones). We can also integrate custom-trained LLMs for specific clients. 
In other words, all of your data is private and not used to train any model.

chatGPT is a simple experience for individuals, but it is not suited for teams. It is not integrated with your tools, you can't upload documents or data for context, and you have to start from scratch on every conversation. Argil allows you to integrate your tools, and create repeatable. By giving chatGPT memory, integrations and making it shareable for your teams, we allow teams to be 30% more efficient.

The experts our clients prefer are:
1) Marketing: can write any type of content in your style, repurpose content (for example youtube videos to linkedin posts) and help you find topics
2) Analyst: feed him with PDFs, webpages & notes and it will analyze a market, startup or client and generate a report (complete with graphs)
3) Assistant: plug him to all the administrative documents of your company and it will help you fill any annoying process such as forms, documents, grants, etc.

Yes, unlike chatGPT, Argil assistants can generate documents (word & excel formats) that you can download directly. This is very convenient for specific use cases, for example when generating outlines or reports.

When writing any type of content, the #1 pain is chatGPT not writing in your style, your format, with your criteria. We have built a specific style trainer that imitates your style and the type of outputs you're looking for. Try it for free!