juanjuandog

juanjuandog / FinSight-AI

Public

AI equity research agent with resilient workflows, Redis Lua single-flight, pgvector RAG, versioned reports, evidence tracing, and RAG evaluation.

ai-agent financial-research llm-evaluation pgvector postgresql
21
1
85% credibility
Found May 20, 2026 at 21 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Java
AI Summary

FinSight AI is an open-source tool that transforms raw financial data into trustworthy stock research reports. It automatically pulls public financial documents and market data for Chinese A-share stocks, calculates key financial health indicators, detects risk signals, and builds a searchable evidence library. The system uses AI to generate structured analysis reports with ratings and confidence scores, and can answer follow-up questions by citing specific documents. Everything runs locally with fallback modes so the analysis keeps working even without external AI services.

How It Works

1
🔍 You discover the project

You find FinSight AI while searching for open-source stock analysis tools. The dashboard preview catches your eye.

2
🚀 You launch everything with one click

A single script starts the dashboard, database, message queue, and AI assistant all at once. Everything is ready in moments.

3
📊 You pick a stock to analyze

You enter a Chinese stock code like 600519 (Kweichow Moutai) and the system immediately shows real-time prices and company info.

4
📄 The system reads financial reports automatically

FinSight pulls public financial statements, annual reports, and announcements from official sources without you lifting a finger.

5
The system works behind the scenes
📈
Financial metrics are calculated

Profitability, growth rates, cash flow quality, and risk signals are computed from the raw numbers.

🔗
A company knowledge graph is built

The system maps company events, industry relationships, and financial connections into a visual timeline.

📚
Documents are indexed for search

Every report and announcement is split into searchable chunks so answers can cite their sources.

6
🤖 You get an AI research report

The AI assistant generates a structured analysis with a rating (positive/neutral/cautious), confidence score, and cited evidence.

7
❓ You can ask follow-up questions

Type any question about the company and get answers backed by actual financial documents, not guesses.

✅ You have a complete research report

You now have a trustworthy analysis with clear ratings, cited evidence, and a record of how every conclusion was reached.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 21 to 21 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is FinSight-AI?

FinSight-AI is an open-source AI equity research agent built in Java with a Python sidecar that analyzes Chinese A-share stocks. The system fetches financial data from public sources, computes metrics, and generates structured investment reports grounded in evidence. It uses Spring Boot for backend orchestration, a FastAPI service for document parsing and embedding, PostgreSQL with pgvector for hybrid retrieval, Redis for caching, and RabbitMQ for async workflows. A static dashboard and demo scripts let you spin up a working prototype in minutes.

Why is it gaining traction?

The project goes beyond typical RAG demos by focusing on production-grade reliability. Its workflow orchestrator handles task failures with retry logic and dead-letter queues, so long-running research pipelines do not silently break. The evidence-traceable RAG shows exactly which documents back each answer, a practical concern for compliance and audit trails. Versioned reports tied to data snapshots prevent stale conclusions from being reused. Built-in evaluation covers hallucination risk and evidence coverage, letting teams catch regressions before shipping. The system falls back to deterministic rule-based analysis when Ollama is unavailable, so local demos and tests stay green without external dependencies.

Who should use this?

Backend engineers building equity research systems who want battle-tested patterns for workflow orchestration, caching, and RAG evaluation. Fintech teams that need source-grounded answers for regulatory reporting. Researchers analyzing A-share stocks who want to wire financial data ingestion into AI-driven report generation without starting from scratch.

Verdict

FinSight-AI fills a real gap for teams that need dependable AI equity research infrastructure, not another toy RAG demo. The credibility score of 0.85% and modest star count signal early-stage maturity, but the architecture and documentation demonstrate thoughtful engineering. Expect to invest in understanding the workflow stages before deploying to production.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.