What it is
Folionix continuously watches the IDX market, values every asset class you hold, and gives you one honest net-worth figure. It runs local LLM analysis on your positions, sends Telegram alerts when something moves, and serves a private Next.js dashboard. Your positions, history, and analysis stay in your own self-hosted Supabase, never a third party.
Features
five things it doesMulti-asset tracking
IDX stocks, gold, mutual funds, and bonds valued together as a single portfolio.
Continuous monitoring
A session-aware orchestrator polls faster during market hours and detects price and volume signals.
Local LLM analysis
Runs analysis through a local model (Ollama). Deep runs vote across investor personas for a consensus call.
Telegram alerts
Actionable alerts on signals, with dedup so you are not spammed by the same move twice.
Private dashboard
A Next.js dashboard with full CRUD over positions, watchlist, gold, funds, and bonds.
Weekly review
An automated week-over-week retrospective with an AI self-critique, emailed and pinged to Telegram.
How it works
data flowFetch
Market data pulled per ticker from Yahoo Finance and Cermati, cached to Supabase.
Analyze
A local LLM scores positions and produces a recommendation for each holding.
Alert
Signals fire Telegram alerts; everything renders on your own Next.js dashboard.
Stack
built withFAQ
common questionsWhat is Folionix?
A self-hosted portfolio analyzer for Indonesian (IDX) markets. It tracks stocks, gold, mutual funds, and bonds in one dashboard, runs local LLM analysis, and sends Telegram alerts. All data lives on your own server.
Which asset types does it support?
IDX-listed stocks, physical gold, mutual funds (reksa dana), and government bonds, valued together as one net worth figure.
Does it send my data to a third party?
No. Folionix is self-hosted. Market data is fetched from public providers, but your positions, analysis, and history are stored in your own self-hosted Supabase and never sent to a third-party service.
Does it require a paid LLM API?
No. It runs on a local LLM (Ollama) by default, with an optional fallback backend. No paid API key is required.
How is it deployed?
With Docker Compose: the LangGraph orchestrator, the Telegram bot, an analysis worker, and the Next.js dashboard. Supabase runs as its own separate stack.