When your stock picker is a large language model

What it does

go-stock is a Wails + NaiveUI desktop application that fetches market data for A-shares, Hong Kong, and US equities, then hands the analysis to an LLM of your choice. It keeps all data local, offers K-line technical indicators, sentiment scoring, and push alerts for price swings. The author is explicit that this is “for entertainment” and that AI-generated investment advice should be treated as educational, not actionable.

The interesting bit

The project treats the LLM as a modular backend: you can swap in OpenAI, DeepSeek, Ollama, LMStudio, AnythingLLM, or several Chinese-hosted inference platforms (SiliconFlow, Volcano Ark, Alibaba Bailian). The UI layer is Go talking to a Vue3 frontend via Wails, which is an unusual stack in a space dominated by Python notebooks and browser dashboards. There is also a sponsorship ladder—free users pull updates from GitHub; paid tiers get CDN delivery and extra AI assistant features.

Key highlights

• Supports A-shares, Hong Kong, and US markets with local data storage
• Pluggable LLM backends including local runners (Ollama, LMStudio) and Chinese cloud APIs
• AI-assisted stock screening, multi-turn chat for follow-up questions, and configurable prompt templates
• K-line charts with technical indicators; recent additions include fund K-lines and top-10 holdings data
• MCP tool-calling support and two agent modes (React, PlanExecute) added in 2026 releases

Caveats

• Development and testing are Windows-centric; macOS support exists but is labeled “untested or functionally limited”
• The README carries heavy sponsorship and referral-link content, including a tiered VIP system for updates and features
• Some roadmap items (ETF full support, stock-analysis knowledge base) remain under construction

Verdict

Worth a look if you want a local-first, LLM-augmented stock screener with Chinese market coverage and don’t mind navigating sponsor links. Skip it if you need battle-tested cross-platform reliability or are allergic to freemium update tiers.