How to use Claude to expand a financial terminal
The author expanded Neuberg from 30 to 516 panels with Claude Code and reviewed and integrated the code. Most newly added panels use seeded or simulated data; the core 30 use real data.
news.ycombinator.com
Prompt to try
Editorial adaptation, not the source author’s exact prompt. Replace bracketed inputs and review tool permissions first.
Help expand [financial dashboard] from [existing panels] to [target coverage]. Generate panel code for [market categories], mark seeded or simulated data clearly, and keep live feeds distinguishable. Present each change for my review and integration, including [charting and trading integrations].
Reach people exploring AI agents
See placements →What happened
Neuberg grew from 30 to 516 panels in about three weeks; the core 30 use real data, while most new panels use seeded or simulated data.
Reported by saratsai on Hacker NewsHow the agent did it
- Built the Neuberg trading terminal with Claude Code.
- Expanded its interface from 30 panels to 516 over about three weeks.
- Reviewed and integrated the AI-written panel code.
Where a person stepped in
The author explicitly says they reviewed and integrated everything, and discloses that most new panels use simulated or seeded data.
A person approved each step before the agent continued.
The author built Neuberg because they wanted a Bloomberg-style terminal but could not justify its cost as an independent trader. They used Claude Code to develop a trading interface spanning fixed income, derivatives, commodities, equities, credit, macro and alternative assets. Its panels can be dragged, dropped and rearranged.
The platform scrapes financial news and applies AI sentiment analysis with first- and second-order impact detection. It includes indicator-based charting, a world map of news events and conflict-zone heatmaps, and trading integrations for Hyperliquid, Polymarket and Alpaca. Alpaca supports both paper and live stock trading in the described terminal.
Over roughly three weeks, the author expanded the platform from 30 panels to 516, mostly through Claude Code. They estimate that AI wrote around 90% of panel code, while they reviewed and integrated everything. The source explicitly limits the apparent coverage: most of the 486 newly added panels use seeded or simulated data. The original 30 panels, including news, charting, trading, calendar and heatmaps, use real data.
The author describes a solo project with rough edges, licensed under BSL 1.1 and free for non-commercial use. They requested feedback on architecture, UI and panel coverage.
Who it's for: Developers prototyping market dashboards who can review generated code and distinguish live feeds from simulated panels.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · saratsai
Frequently asked questions
Are all 516 panels connected to live data?
No. The author says the core 30 panels use real data, while most of the 486 newly added panels currently use seeded or simulated data and are not yet connected to live feeds.
Who reviewed the AI-written code?
The author says AI wrote around 90% of the panel code and that they reviewed and integrated everything during the roughly three-week expansion.
More details
About this account
Claude Code generated most panel code; the author estimates AI wrote around 90% during the expansion.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=47596655. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: saratsai on Hacker News
Source published 1 Apr 2026
Source checked 14 Oct 2026
Tools and topics
Claude Code · Hyperliquid · Polymarket · Alpaca
Financial dashboards · Human review · Simulated data
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.