Practical overview
When it fits
Use it when company analysis should connect structured financial information with filings and events, while preserving the report's underlying data snapshot and evidence path.
Who it is for
- Researchers investigating A-share companies through market and financial data.
- Developers exploring recoverable AI research workflows and hybrid retrieval.
Documented uses
- Inspect a company's quote, historical closing-price curve and key financial metrics.
- Generate a structured analysis with a conclusion, confidence, supporting factors and risk factors.
- Search filings, announcements and structured metrics for evidence.
- Review company disclosures, metric changes and risk signals on a timeline.
- Keep a watchlist of companies for follow-up research.
Requirements & limits
The repository is public and declares an MIT license. The documented setup options are a Java 17 and Maven lightweight preview or a Docker Compose full stack.
- The lightweight preview uses in-memory adapters; the full workflow and retrieval stack requires Docker Compose and its supporting services.
- The full Compose stack includes a FastAPI AI sidecar alongside the Spring Boot backend, PostgreSQL/pgvector, Redis and RabbitMQ.
- Reports retain versions and data-snapshot identifiers, and retrieval records an evidence trace.
- The project describes its output as research aid rather than investment advice and does not present itself as an automated trading system.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Start a lightweight local preview
Run the documented backend preview using local in-memory adapters, without infrastructure services.
Before you start
- Java 17
- Maven
Steps
Clone the FinSight AI repository.
Change into the backend directory.
Run the Spring Boot application with Maven.
Open the local application at http://localhost:8080.
Expected outcome
A local preview of the product and its core flow.
Guide limitations
- This mode uses in-memory adapters and does not require infrastructure services; it is distinct from the full research stack.
Start the full research stack and demo
Bring up the documented Compose stack and run its quick demo.
Before you start
- Docker Compose
- About 8 GB of free memory for the complete stack
Steps
Clone the FinSight AI repository.
From the repository root, start the services with Docker Compose and build.
Run the quick demo script.
Expected outcome
The Compose stack includes PostgreSQL with pgvector, Redis, RabbitMQ, the Spring Boot backend and the FastAPI AI sidecar, followed by the repository's quick demo.
Guide limitations
- The default demo requires no API key and uses Ollama as the default local provider.
- OpenAI-compatible and Anthropic adapters are also described; if a selected model is unavailable or unconfigured, deterministic fallbacks are used.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 9 October 2026
- Original website title
- juanjuandog/FinSight-AI · Official homepage title · retained in its source language