Autotrader – paper trading AI agent for Indian equities
A paper trading AI agent named Autotrader has been tested on Indian equities, specifically the Nifty 500 index. The experiment, which ran for two weeks, concluded with an 8.05% gain on the initial capital of Rs 100,000. The project is paused, but the repository is available for others to fork and continue the work.
- ▪The Autotrader experiment utilized real transaction costs and a live price feed without using real money.
- ▪The trading agent was designed to edit its own strategy based on performance during the trading loop.
- ▪The project included detailed documentation and setup instructions for users interested in replicating the experiment.
Opening excerpt (first ~120 words) tap to expand
autotrade A paper-trading experiment on Indian equities (Nifty 500). Claude itself runs the trading loop on a free GCP VM and edits its own strategy between polls. Real Indian transaction costs (STT, GST, stamp duty, brokerage), real Kite Connect price feed, no real money. Status: paused. Two-week run ended May 8 2026 at Rs 108,049 (+8.05% on Rs 1,00,000 starting capital). The repo is left in a state forkable for anyone who wants to continue or reuse the harness. The full writeup (sessions, incidents, what the experiment actually taught) is at https://www.akashtandon.in/autotrader/. What's in here live.py - single-shot polling loop. Fetches prices, runs strategy, executes paper trades, persists state. strategy.py - trading signals + position sizing.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at GitHub.