simonlin1212
Vibe-Research
Vibe-Research is an open-source personal AI investment research dashboard that integrates A-share, US, and Hong Kong stock market data with a clean UI and a pluggable AI interface, never recommending stocks.
AI repository intelligence
FR-AI / ANALYSISWhy it is worth attention
It addresses the common shortcoming of AI-generated summaries by preserving source-grounded signals, separating evidence from inference, and making uncertainty explicit, all while being bilingual and open source.
Who it is for
- Market researchers seeking transparent analysis
- UX researchers and product managers
- Developers interested in evidence-aware research tools
- Privacy-conscious analysts using public data
Use cases
- Building research panels from public market signals
- Generating de-identified persona cards for user analysis
- Answering research questions with evidence-aware prompts and consensus summaries
- Localizing research reports and caching results for different audiences
Strengths
- Evidence-grounded workflow that separates direct evidence from inference
- Built-in privacy measures including de-identification and masking
- Full English/Simplified Chinese localization with cached reports
- Open source (Apache-2.0) with clear React/FastAPI architecture
Considerations
- Prototype status not yet ready for production use
- Requires user-provided API credentials for full data connector and model-backed runs
- Demo data must be sanitized and no auto-loading of .env files
README quick start
Market Pilot
Market Pilot is an evidence-grounded market research prototype that turns public market signals into traceable research workflows.
Why it exists
AI can summarize large amounts of market conversation quickly, but a polished summary is not the same as inspectable evidence. Market Pilot explores a stricter workflow: preserve source-grounded signals, separate direct evidence from inference, mask personal identifiers, and make uncertainty visible in the final report.
Current capabilities
- Build research panels from collected public signals.
- Generate de-identified persona cards for analysis.
- Route and answer research questions through evidence-aware prompts.
- Summarize consensus, disagreement, limitations, and supporting signals.
- Switch the complete public and authenticated experience between English and Simplified Chinese.
- Read cached report localizations without collecting data or rerunning research.
- Provide a React product interface backed by a FastAPI service.
This repository is a product prototype. Demo data must be labeled as sanitized, and data connectors or model-backed runs require your own authorized provider credentials.
Architecture
- Frontend: React 18, Vite 6, Tailwind CSS 4
- Backend: FastAPI
- Local routing: the frontend calls
/api; Vite proxies requests tohttp://localhost:8001 - Research pipeline: collection, masking, panel construction, evidence-aware answering, aggregation, and report rendering
- Localization: unprefixed English public URLs,
/zhChinese public URLs, account-aware locale resolution, and a validated report-localization cache
See docs/architecture.md for the component map.
Local setup
Requirements: Node.js 24+, npm 11+, Python 3.13+, and Git.
git clone https://github.com/Dgeloe4-yb/market-pilot.git
Set-Location market-pilot
npm.cmd ci
python -m venv .venv
.\.venv\Scripts\python.exe -m pip install -r collector\requirements.txt
Review .env.example, then export only the variables you need into the process environment. The application does not auto-load .env files.
Start the API:
Set-Location collector
..\.venv\Scripts\python.exe -m uvicorn api.app:app --port 8001
Start the frontend from the repository root in another terminal:
Description
Evidence-grounded market research prototype with traceable AI workflows.
Related repositories
Similar projects matched by category, topics, and programming language.
react
Featured
react
React is a JavaScript library for building user interfaces with declarative, component-based, and cross-platform design.
7-e1even
Featured
learn-agent
A collection of notes on building coding agents, derived from the production development of the desktop agent Reina, with each mechanism simplified into a zero-dependency, single-file Node.js demo.