Jinling AI

What is Jinling AI at the moment?

Jinling AI’s Chinese version has indicated that its services are being upgraded, with all functions being transferred to LangAlpha, which is part of the Ginlix portfolio.

LangAlpha is an intelligent agent for financial research, which uses real-time data, public documents, and code execution to carry out research tasks.

Relationship between brand and product

Multi-agent financial research

LangAlpha breaks down research tasks into multiple specialized agents, which then combine their efforts to produce analytical results that can be reviewed.

• Process market data, financial reports, and web content simultaneously.
• Write code and compute data in a sandbox.
• The research plan is approved or modified by the user.
• Adjustment commands can continue to be sent while it is running.
• For each key conclusion, the source is tried to be retained.
• The final decision regarding investment remains the responsibility of the user.

Scope of financial data

• Real-time updates on U.S., A-share, and Hong Kong stock markets.
• Company fundamentals and market performance.
• SEC’s 10-K, 10-Q, and 8-K filings.
• Text transcript of the financial results call and institutional holdings.
• US stock option chains and intraday quotes.
• U.S. Treasury yield curve and macroeconomic data.
• News, web pages, and emotional tagging information.

Publication of research results

Agents can generate structured research documents based on tasks, rather than merely returning a chat-based response.

• Generate a DOCX research memo with citations.
• Create XLSX valuation and financial models.
• Generate a PPTX presentation.
• Format PDF research reports.
• Create interactive charts and real-time dashboards.
• Retain the calculation scripts and data invocation records.

Persistent workspace

A separate workspace can be created for each research goal, to store conversations, files, notes, and long-term context.

• Categorize the items by company, industry, or investment rationale.
• Upload PDF and Markdown research materials.
• Save the models and reports generated by the agent.
• The existing context is continued across conversations.
• Different work areas are isolated from one another.

Automation and price triggers

Users can schedule regular analyses, or initiate a full task when stocks or indices reach a target price.

• Create a pre-market briefing for the disk.
• Arrange a weekly review of positions held.
• Set research tasks before and after the financial report is released.
• Trigger analysis based on price or percentage change.
• The results can be sent to the connected channels.
• Automatic analysis does not execute transactions automatically.

Channel and model integration

• The managed version supports messaging platforms such as Slack.
• You can bring your own model keys such as Claude and GPT.
• Different models can be used for deep or fast tasks.
• Third-party models process data in accordance with their own terms.
• The key should be stored in a dedicated safe, rather than in the prompt.

LangAlpha usage guide

1. Enter the current LangAlpha hosting platform.
2. Register an account and create a research workspace.
3. Explain the market, the target, and the research objectives.
4. Upload research materials for which usage rights are available.
5. Check the execution plan proposed by the agent.
6. Approve, modify, or reduce the scope of the research.
7. Allow multiple agents to collect and compute in parallel.
8. Check the data timing and the source of reference.
9. Check the valuation assumptions and calculation formulas.
10. Choose an output format such as DOCX, XLSX, or PDF.
11. Save the results and set up subsequent automated tasks.
12. Verify once again using an independent source before making the transaction.

Price of the managed version

The hosting platform can be used for free; the paid plans increase the monthly usage limit and allow for additional credit top-ups.

The above information was verified on August 30, 2026; the limits, amounts, taxes, and availability by region are subject to those indicated on the settlement page.

Open source and APIs

The core project of LangAlpha is licensed under the Apache-2.0 license, and its Web, backend, plugin, and API documentation are made available publicly.

• The source code repository includes the front-end, back-end, and command-line interface.
• Provides APIs for chat streams, workspaces, and process status.
• Interactive API documentation can be accessed after running it locally.
• The managed version of data services does not mean that everything is open source.
• Third-party data and models are still subject to their respective licensing restrictions.

Self-hosted tutorial

1. Prepare Docker and sufficient server resources.
2. Obtain the official LangAlpha source code.
3. The runtime configuration wizard is used to generate the environment configuration.
4. Select the model provider and use an existing key.
5. Configure market data, search, and sandbox services as needed.
6. Start the database, cache, backend, and frontend.
7. Check the health status and interactive interface documentation.
8. Create an isolated workspace to carry out the testing.
9. Configure backups, encryption, and access control before going live.

Self-hosting costs

• Open-source licenses do not charge any fees for the software.
• Calling or subscribing to large models may incur costs.
• Real-time market data and professional information may be subject to fees.
• Cloud sandboxing, search, and scraping services may incur fees.
• The isolation capability of the local Docker sandbox is reduced.
• It is also necessary to calculate the costs for servers, storage, and maintenance.

Privacy and data processing

The managed version states that it will not use user conversations to train AI, nor will it sell personal data.

• The account stores the name, email address, and authorization details.
• Conversations, files, and outputs are saved in the workspace.
• The request will be sent to the selected third-party model.
• The code is executed in an isolated sandbox.
• Workspaces and sessions can be deleted by users.
• In principle, the deletion of a complete account is processed within 30 days.

Financial risks and fact verification

• There may still be delays or gaps in the real-time market data.
• Public documents may be misinterpreted by the model.
• Valuation relies heavily on assumptions and input data.
• News sentiment cannot directly indicate the direction of prices.
• Automatically triggering a study does not mean automatic trading.
• The content provided by the platform does not constitute investment advice.

Which users are it suitable for

• Investors who require in-depth company research.
• Analysts who create financial models.
• Researchers who track financial reports and SEC documents.
• Teams that wish to have market briefs generated automatically.
• Technical organizations that require local deployment.
• Open-source contributors to the development of financial agents.

Product advantages

• Combine financial data, code, and agents.
• The research conclusions emphasize the source and the calculation process.
• It is possible to generate various editable output files.
• It supports long-term workspaces and session-crossing memory.
• Provides tasks triggered by schedule and price.
• The core code can be self-hosted and extended.

Usage restrictions and precautions

• The San Francisco AI site has been migrated to the new product.
• The specific payment amount for the managed version has not yet been made public.
• Professional data and models may incur additional fees.
• Self-hosting requires deployment and security capabilities.
• Third-party models handle the query context.
• Any trading decision must be independently verified.

Frequently Asked Questions

Can Jinling AI still be used?

The original Chinese version indicated a service upgrade; all features have been moved to the LangAlpha platform, which is part of Ginlix.

What is LangAlpha mainly used for?

It utilizes multi-agent systems, real-time financial data, and code execution to carry out tasks related to research, modeling, reporting, and generating charts.

Does LangAlpha trade automatically?

No, it can initiate research based on time or price, but the decision to proceed with a transaction remains up to the user.

Is LangAlpha free?

The managed version can be started for free, while the paid version increases the monthly quota; for the open-source version, you have to cover the costs related to models and data on your own.

Is LangAlpha an open-source project?

Yes, the core projects are licensed under the Apache-2.0 license, but the hosting data services and third-party data are not all open source.

Will LangAlpha use conversations to train the model?

The privacy policy for the managed version states that AI is not trained using conversations; instead, queries are processed by the selected third-party model.

Can the analysis provided by LangAlpha be used as investment advice?

No, the platform is designed as a research tool; all data, hypotheses, and conclusions must be verified independently.