Roadmap & Changelog

Better statistics, composite scoring for portfolios and models, and clearer model-level vs. portfolio-level metrics.

Talk with any strategy model about any stock, with real-time ratings in the loop.

Surface daily ratings and signals in a way that fits how you scan the market.

Additional prompt strategies, multi-agent rating systems, and different underlying models.

Try unconventional engines—for example swarm-style prediction inspired by multi-agent simulation research. MiroFish (reference)

Let power users define and iterate on their own rating strategies inside the platform.

Compare models side by side with fair, transparent ranking.

Go beyond presets so portfolio rules match how you actually invest.

More customization, real-time signals, and other improvements to what hits your inbox.

Help users understand buy/sell tax implications when choosing portfolios and rebalances.

Potential integration with brokerage APIs so trades can be placed on your behalf—where regulations and partners allow.

First production cut: portfolio configs on top of the AI rating engine—Explore Portfolios, Your Portfolios, and per-portfolio performance metrics. Performance tracking across the platform, plus the rating execution pipeline (cron, batch runs, config-scoped performance).

Public navigation and page structure refresh; platform tabs and remade performance and strategy model experiences; subscription tiers and premium stock access; email/password auth with forgot/reset; ongoing speed and UI polish.

Portfolio config layer atop the AI rating layer, weekly strategy performance framework, schema and cron hardening, RLS and account sync improvements, and earlier AI analysis pipeline work.