DEDUPEDWho it is forAn investor who prefers asking follow-up questions in a browser over reading a fixed report.
The problemA one-shot report cannot answer your next question about the same stock.
The outcomeA local workbench with history, full reports, backtests, holdings and an agent you can question in several turns.
What you need first
- A GitHub account (for the Actions route) or Python and pip locally
- At least one AI model API key, such as Gemini, Anthropic, an OpenAI compatible key, or Ollama locally
- At least one notification channel such as a Telegram bot, Discord webhook or email
- Optional: a news search API key, which the README says strongly improves news and catalyst quality
Steps
- Configure .env with any working AI key
- Start the workbench with
python main.py --webui - Open http://127.0.0.1:8000 in your browser
- Use the /chat page to ask about a stock with one of the built-in strategies, and follow up in several turns
- Set AGENT_MODE=false if you want to turn the agent off
Prompt for your agent
Run ZhuLinsen/daily_stock_analysis with python main.py --webui, open the /chat page and ask about <ticker> using the trend strategy. Summarise the answer and list which data tools the agent called.
Caveat. Custom strategy files and multi-agent mode are marked experimental. Not financial advice: paper trade first, and remember you can lose money.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for ZhuLinsen/daily_stock_analysis and "Ask an AI agent about a stock in a web workbench"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer