About NightIndex

What NightIndex is

NightIndex is an AI-assisted market-intelligence site. It continuously collects items from public sources, organizes them into eight desks, publishes daily digests in Chinese and English, and attributes every claim to its source.

It is not a news organization and does not give investment advice.

How a digest is made

1. Collection. Titles and excerpts are collected continuously from 60+ public sources: official bodies, exchanges, major media, industry associations and public research pages.
2. Classification. Each item is assigned by semantic similarity to exactly one desk; items that fit none are dropped.
3. Three filters. A relevance threshold, cross-source de-duplication that keeps one item per event, and a model review that removes ads, off-topic and low-quality items.
4. Writing. The filtered items for a desk and time window are given to a large language model. The model may only use those items and nothing outside them; every numbered [n] in the text maps to an entry in the citation list.
5. Publishing. Two editions are published each day—an overnight brief and a day digest—in Chinese and English.

Source policy

• For third-party items, we show only the title, an excerpt of at most 160 characters and a link to the original; no full text is republished.
• Paid-subscription research, including broker and bank reports, is excluded from public digests and the home-page wire.
• Market and positioning data are never published as tables, charts or downloads.
• Citations are numbered in reading order with the source name, original title and publication time.

Corrections and withdrawals

• A digest may be regenerated when sources update or a model error is found; the URL stays the same and the displayed time is updated.
• Withdrawn digests return HTTP 410 with a notice rather than vanishing.
• If you spot an error, contact us via your account page on the site.

Citing NightIndex

Cite as “NightIndex” with a link to the corresponding digest page. Machine-readable entry points are /llms.txt, /sitemap.xml and the RSS feeds.