myai 2.0

A lot has shipped in a short window. Here is where the platform stands.

A real foundation

myai now runs on a single relational backbone. Everything you create lives in one consistent data model, reachable through one clean query surface that respects who is allowed to see what. The day-to-day payoff is speed and consistency: workspaces that used to take close to thirty seconds to open now load in under a second, recent items reflect real activity, and long conversations stay responsive.

Build an app in the conversation, run it anywhere

Apps are now real, multi-file projects. Ask myai to build one and watch it render live as it takes shape, with working data behind it, before you ever publish. When you want to keep going on your own machine, download the app as a runnable folder, edit it in your own tools, and publish it back. Apps can even import each other, so a useful piece becomes a building block.

Share what an app does, not the keys behind it

A shared app runs with its owner's authority. You can hand a teammate a dashboard that reads a connected system without handing over the credentials to that system. They get everything the app can do, and nothing it was not meant to expose.

Your choice of model, always on

Choose how your work runs from the chat panel, or let myai choose. When a provider has a rough moment, myai moves to another capable model automatically and tells you it switched, so a busy provider no longer means a dead turn. Leading models now run across more than one cloud for exactly this reason.

Query everything you have

Ask questions across your work the way you would a database: search the full text inside your PDFs, read back the turn-by-turn history of a job, look people up by email. Results come back complete, with no silent caps.

Reliability: search enrichment now self-heals, dashboards and the share dialog open instantly, and long conversations stay fast.