Documentation

Help & Documentation

Everything you need to deploy autonomous AI agents, interpret intelligence signals, and get the most out of DiffiWeb.

01

Getting Started

1

Create an Account

Visit diffiweb.com and click "Get Started". Register with your email and a secure password. A verification may be sent to confirm your address.

2

Log In

Enter your credentials at diffiweb.com/login. Your session is secured with an encrypted cookie — you stay logged in for 7 days.

3

Open the Dashboard

After login you land on the Intelligence Hub. Use the sidebar (desktop) or the bottom tab bar (mobile) to switch between the 11 intelligence modules.

Navigating on Mobile

  • • Tap the ☰ hamburger (top-left) to open the full module drawer.
  • • The bottom tab bar gives one-tap access to Hub, Signals, Stocks, Vault, and More.
  • • The coloured FAB (floating button above the tab bar) launches the action for the current module — no need to open the drawer.
  • • Tap More in the tab bar to see all 11 modules in a grid sheet.
02

Intelligence Modules

Intelligence Hub

Core

The central command centre. Run autonomous agents that crawl government portals, procurement databases, and news feeds to surface tenders, RFPs, and grants in real time.

How to Use

  1. 1Click Run Agent (desktop sidebar) or the amber FAB on mobile.
  2. 2Paste a target URL — a government portal, eProcurement site, or news page.
  3. 3Enter a search query e.g. "infrastructure tender Maharashtra 2026".
  4. 4Choose the number of agents (1–4). More agents = broader parallel coverage.
  5. 5Set the item limit (max 20) and hit Launch Agent.
  6. 6Watch the live log stream. Extracted items land in The Vault automatically.

Pro Tips

  • Stagger multiple agents across different state portals for national coverage.
  • Use specific queries — "civil construction contract ₹50 crore" beats "tender".
  • Agents run for up to 10 minutes; keep the tab open or check The Vault after.

Stocks Intelligence

Live Data

Real-time stock quotes from NSE/BSE via Yahoo Finance. Track a personal watchlist and get AI-powered Alpha Bucketing (Bullish / Neutral / Bearish) signals.

How to Use

  1. 1Go to Stocks Intelligence from the sidebar or bottom tab bar.
  2. 2In the Watchlist panel on the right, type a ticker (e.g. RELIANCE, TCS, INFY).
  3. 3Hit Add — quotes refresh every 30 seconds automatically.
  4. 4The main panel shows the Alpha Intelligence feed with AI-scored signals.
  5. 5Remove a ticker by clicking the ✕ next to it in the watchlist.

Pro Tips

  • Use NSE tickers without .NS suffix — the system appends it automatically.
  • Alpha Bucketing scores are refreshed each time an agent run completes.
  • Combine with Financial Signals for full market intelligence coverage.

Financial Signals

Market Intel

Deploy agents against financial news portals, regulatory filings, and market data sources. Each item is scored 1–10 for market impact.

How to Use

  1. 1Click New Signal Engine (sidebar or FAB).
  2. 2Enter the target URL — e.g. moneycontrol.com, economictimes.com, SEBI filings.
  3. 3Enter a query like "RBI rate decision 2026" or "NIFTY earnings surprise".
  4. 4Set agent count and launch.
  5. 5Results appear sorted by Impact Score (highest first).

Pro Tips

  • Impact Score ≥ 8 = high volatility signal. Set up multiple agents for breadth.
  • Target earning-season dates for maximum signal density.
  • Results are saved to The Vault with timestamps for historical review.

Competitor Tracker

Strategy

Monitor competitor websites, press releases, LinkedIn pages, and product updates. Each item is scored by strategic impact.

How to Use

  1. 1Click New Competitor Intel (sidebar or FAB).
  2. 2Enter the competitor's website or a relevant news aggregator.
  3. 3Query with specific terms: "product launch", "funding round", "partnership".
  4. 4Launch 1–4 agents. Each result gets an Impact Score and competitive insight.

Pro Tips

  • Track 3–4 core competitors simultaneously using multiple agents.
  • Set queries around their product names for precise signal extraction.
  • Combine with Web Scrapers to pull data from LinkedIn or Twitter/X.

Vulnerability Scanner

Security

Autonomous security agents crawl your target systems and public CVE databases, extracting vulnerabilities with severity scoring (CVSS-style 1–10).

How to Use

  1. 1Click New Scan (sidebar or FAB).
  2. 2Enter the target — a URL, IP range description, or CVE database page.
  3. 3Add a query: e.g. "SQL injection", "authentication bypass", "CVE-2025".
  4. 4Launch. Agents extract vulnerability details, severity scores, and remediation hints.
  5. 5High-severity items (score ≥ 8) are flagged in red in the results.

Pro Tips

  • Only scan systems you own or have permission to test.
  • Use NVD (nvd.nist.gov) or CVE.mitre.org as targets for threat intelligence.
  • Cross-reference with your tech stack for targeted remediation priority.

Supply Chain Radar

Global Intel

Monitor global supply chain disruptions, port delays, trade news, and procurement anomalies across logistics portals and government databases.

How to Use

  1. 1Click New Global Radar (sidebar or FAB).
  2. 2Target a logistics portal, trade news site, or government trade database.
  3. 3Query with route-specific or commodity-specific terms.
  4. 4Results include disruption signals, lead-time anomalies, and supplier news.

Pro Tips

  • Target IHS Markit, Freightos, or customs portals for best signal quality.
  • Run weekly to maintain an up-to-date disruption map.
  • Filter results by country or commodity in The Vault using Semantic Search.

Web Scrapers

Custom

General-purpose scraping agents for any URL. Configure a custom extraction profile and optionally receive results via email.

How to Use

  1. 1Click New Web Scraper (sidebar or FAB).
  2. 2Enter any target URL.
  3. 3Enter an extraction query — what data you want pulled.
  4. 4Optionally enter an email address to receive results as a report.
  5. 5Set limit and launch. Results appear in The Vault.

Pro Tips

  • Works on most public pages — news articles, directories, marketplaces.
  • Use the email delivery option for async scraping jobs you don't need to watch live.
  • Combine multiple scrapers for comprehensive market research in one session.

Medical Intel

Research

Autonomous agents crawl PubMed, arXiv (q-bio), and Nature to surface peer-reviewed research papers with structured metadata.

How to Use

  1. 1Click New Paper Scan (sidebar or FAB).
  2. 2Choose the source: PubMed, arXiv, Nature, or a custom URL.
  3. 3Enter your research query: e.g. "CRISPR cancer therapy 2025".
  4. 4Launch. Papers are extracted with title, authors, abstract, date, and link.
  5. 5Results are saved to The Vault for citation and reference.

Pro Tips

  • Use MeSH terms on PubMed for precision: "neoplasms[MeSH] AND immunotherapy".
  • arXiv is better for pre-prints; PubMed for peer-reviewed clinical research.
  • Run recurring scans weekly on the same query to track a research area over time.

The Vault

Storage

All extracted intelligence is automatically saved here. Browse, filter, and delete your entire research history across all modules.

How to Use

  1. 1Navigate to The Vault from the sidebar or bottom tab.
  2. 2All previously extracted items appear as cards sorted by date.
  3. 3Click the external link icon on any card to open the original source.
  4. 4Click the trash icon to delete individual items.
  5. 5Use Semantic Search for AI-powered querying across all saved data.

Pro Tips

  • The Vault is scoped to your account — no other user can see your data.
  • Data is stored in a vector database, enabling semantic (meaning-based) search.
  • Regular cleanup keeps your search results more relevant.
03

Frequently Asked Questions

How many agents can I run at once?

Up to 4 simultaneous agents per module. You can run agents across multiple modules concurrently — e.g. 4 Tender agents + 4 Vulnerability Scanners at the same time.

Where is my data stored?

All extracted intelligence is stored in Upstash Vector (a vector database) and Upstash Redis. Your data is isolated by your user account — no other user can see or access it.

How long does a scraping run take?

Most runs complete in 2–8 minutes depending on the target site's complexity and the number of agents. The live log stream updates in real time so you can monitor progress.

Can I scrape any website?

DiffiWeb agents can access most publicly available websites. Respect the terms of service of websites you target. Do not use DiffiWeb to scrape private, authenticated, or restricted systems without authorisation.

Why are my stock quotes showing a delay?

Quotes are sourced from Yahoo Finance's market data API and refresh every 30 seconds. During pre-market or after-hours sessions, you may see the previous close price.

I ran an agent but nothing appeared in The Vault.

This can happen if the target page had no extractable items matching your query, or if the site blocked the agent. Try a different target URL or a broader query. Check the live log for "0 items extracted" messages.

Is my data deleted if I close the browser?

No. All data is persisted in the cloud database. Your Vault items remain intact until you manually delete them.

How do I reset my password?

On the login page, click "Forgot password?" and enter your email. A reset link will be sent via Resend email. Check your spam folder if it doesn't arrive within 5 minutes.

04

Support

Need help with something not covered here?

Reach out via email and we'll get back to you within 24 hours on business days.

swapwarick@diffiweb.com