Privacy Policy

Agent Adda is a research-grade equities and derivatives intelligence platform built for the National Stock Exchange of India (NSE). Its editorial principle is simple: Hold. Think. Then act. Genuine research requires investigation — understanding market context, selecting appropriate data, gathering evidence, and conducting comparative analysis before forming any conclusion.

The platform processes 137 indices, evaluates 48 investment prospects across 12 analytical dimensions per holding, and routes every question through a five-layer architecture: intent parsing → data retrieval (local-first, with public API fallbacks) → evidence gathering via specialised tools → LLM synthesis → persistent report generation.

Talk 2 Stocks is the conversational web interface to this engine. You ask questions in plain English or use slash commands; the system selects the appropriate data window, runs the relevant tools, and returns structured, sourced research — not ephemeral chat responses. The Agent Adda CLI offers the same engine with additional depth for backtesting, voice briefings, sector rotation tracking, and PDF reports.

The platform surfaces 12 intelligence types:

• —Natural-language queries with explicit data-mode control (live / EOD / auto)
• —Intraday scanner — nine scanning approaches on 15-minute charts
• —EOD screeners — ten end-of-day filtering strategies across ~1,500 NSE stocks
• —Background monitors — fourteen continuously-running detection mechanisms
• —Watchlist alerts — price and momentum triggers on NSE symbols
• —Deep search — parallel investigation across 11 information categories
• —Document analysis — PDF and spreadsheet reading with 360° stock evaluation
• —Forensic accounting — three financial health screening models
• —CANSLIM evaluation — William O'Neil's growth quality methodology
• —Options and F&O — live chains with strategy builders
• —RIC investigations — eight named multi-step research templates (Sherlock, Breakout Hunter, Earnings Playbook, Sector X-Ray, and more)
• —Report generation — HTML, PDF, and Markdown output formats

The scoring engine synthesises three investment frameworks: Stan Weinstein's four-stage cycle, William O'Neil's CANSLIM criteria, and Mark Minervini's trend templates — implementing a top-down flow from macro regime detection through sector rotation to individual stock stage classification and execution timing.

Agent Adda is a research and learning tool, not a trading recommendation engine. All outputs are educational and informational. No analysis constitutes a recommendation to buy, sell, or hold any security. Agent Adda is not registered with SEBI as an investment advisor.

• —Session identifiers — a random ID generated in your browser when you open Talk 2 Stocks. It is not linked to your name, email, or device fingerprint.
• —Conversation turns — your questions and the answers returned within a session are held in server memory only for the duration of that session (up to 1 hour of inactivity). They are not written to a persistent database or log file.
• —Watchlist symbols — stored only in your browser's localStorage. They never leave your device unless you include them in a question.

We do not collect your name, email address, phone number, financial positions, brokerage account details, or any other personally identifying information.

Each browser session gets its own isolated conversation context. Sessions are evicted from memory after 1 hour of inactivity — after which no record of the conversation is retained on our servers.

When you send a question that requires AI synthesis, the question text and relevant market data are sent to OpenAI's API to generate a response. See the Third-party services section below for details.

Market data requests (prices, filings, breadth) go directly to NSE's public APIs and Screener.in. Your question text is not forwarded to those sources.

Agent Adda depends on the following external services:

Talk 2 Stocks uses localStorage to remember your watchlist and UI preferences (e.g. collapsed sidebar) between visits. This data stays in your browser. You can clear it at any time via your browser's "Clear site data" setting — it has no effect on our servers.

We do not use tracking cookies, analytics pixels, or cross-site identifiers of any kind.

Conversation sessions — evicted from server memory after 1 hour of inactivity. No persistent copy is made.

Market data caches — price, earnings, and breadth data are cached server-side for 18–24 hours to reduce API load. These caches contain only publicly available market data — no user information.

Server logs — our web server may log standard HTTP request metadata (timestamp, endpoint, response code, approximate request size). IP addresses, if logged, are not linked to conversation content and are rotated within 7 days.

• —Clear your session — close the browser tab. The session expires within 1 hour and is purged from server memory.
• —Clear local data — use your browser's site data settings to remove localStorage entries for agentadda.in.
• —Opt out of AI synthesis — if you prefer answers from deterministic rules only (no OpenAI calls), set the environment variable T2S_USE_AGENT_BRIDGE=0 when running a self-hosted instance.

If we make material changes to how we handle data, we will update the date at the top of this page. For significant changes we will post a notice on the Agent Adda platform home page.