nokris.com

What It Is

nokris describes itself in the page title as an “LLM-style price-action predictor,” meaning an AI tool for price-action forecasting, and mentions the use of “quantile + codex models.” Judging from the name and description, it may focus on financial market price movements, trading signals, or probabilistic forecasting in quantitative research. However, the current page only provides a login entry point and clearly states that “access is restricted to allow-listed GitHub accounts,” which means the product is not publicly available.

Core Capabilities and Models

The only technical information that can be confirmed from the public page is “LLM-style,” “quantile,” and “codex models.” This suggests it may use a modeling approach similar to large language models, combined with quantile-based prediction to represent uncertainty in price movements. However, the page does not disclose training data sources, supported asset classes, forecast horizons, backtesting metrics, sample outputs, or model update frequency, so its actual predictive performance and stability cannot be assessed.

Pricing, Trial, and Integrations

The page does not provide information about free quotas, trial policies, subscription pricing, or enterprise plans. Login is handled via GitHub but restricted to allow-listed accounts, meaning regular users cannot sign up and try it directly. The page also does not mention APIs, SDKs, webhooks, exchanges, brokers, TradingView, or other integrations, so it is currently unclear whether it is suitable for automated trading or quantitative workflow integration.

Pros and Cons

Its main advantage is a very focused positioning around the specific use case of price-action prediction. If the “quantile” approach is implemented well, it could theoretically provide probability ranges rather than single-point forecasts. The drawbacks are equally clear: very limited public information, restricted access, no pricing or privacy details, and no backtests or sample validations. For a financial forecasting tool, the lack of transparent metrics significantly increases the difficulty of evaluation.

Who It’s For and Access from China

nokris is better suited to quantitative researchers, financial AI developers, or early test users who already have allow-list access. For ordinary traders, enterprise teams, or users in China, its practical usability is currently difficult to evaluate. Access from China cannot be determined from the public page alone, and supported payment methods are also unknown. If you need immediately usable alternatives, consider TradingView, QuantConnect, Numerai, or Chinese quantitative research platforms such as 聚宽 and 米筐.