Your mirror to the world
DARPAN stands for Distributed, Autonomous, Realtime, Profile, Acquisition, Network. Avatars and bots run across VMs, containers and proxies; schedulers and crawlers run without an analyst at the controls; a new post, breach or connection is processed and notified as it happens. Entity and topic extraction is stored for the long term, so an analyst can pivot months later from the same identifier.
Connect anything, extract everything
DarpanX reads structured and unstructured sources: SQL and NoSQL databases, spreadsheets, Word, PDF and PowerPoint files, HTML, XML and JSON, images, video and audio, and online registries, lists and APIs by scraping or integration. Entity processing extracts names and aliases, contact details, social ids, addresses, organisations, dates, locations and IP addresses. Topic processing finds keywords and concepts with an NLP pipeline and a document comparator. Sentiment analysis scores attitudes and emotion per entity and detects behavioural change over time.
Search the open web, watch the dark one
Live Search queries a selector against a continuously refreshed breach corpus, with every query logged and attributed to a case. Dark web footprints show an organisation's exposure across domains, emails, IPs, usernames, names, phones and passwords. The scheduler runs recurring, delta-aware sweeps.
One name, every platform
Manage Entity resolves a subject across Facebook, Instagram, X, LinkedIn, WhatsApp and Telegram, then pivots to posts, messages, friends and timeline changes. Multi-platform profile search reports status per platform per attempt, and similar-profile matching on handle, content or imagery surfaces impersonation and sock puppets.
Posts are evidence, behaviour is the signal
Observable content is scored against a multi-signal matrix covering hate speech, narcotics, extremism, threats, explicit imagery and more, filterable by source, date and media type. Sentiment overlays the matrix, and a timeline records every profile change across every monitored source.
Every artefact, read
Document analysis extracts entities from Word, PDF, PowerPoint and structured files. Multilingual image OCR embeds extracted text alongside the image and can integrate third-party OCR engines. Frame-aware video analysis and YouTube search turn long-form media into time-coded hits, and the CDR Analyzer turns call-detail records into a summary with key findings.
From a lead to a defensible record
Suspects, evidence, searches and analyses live inside a case. Evidence carries collector, category, location and timestamp; the journey summary logs every transaction; a Person Report with entity details, posts, network, geolocation and sentiment is generated on demand for handover.
Avatars, bots and tracers
Self-sustained avatars with a biography, posting cadence and friends list act as honeytraps where passive collection fails. Schedule-driven bots crawl social platforms, registries and forums, with dark web bots when needed and a Bots Manager for health and throughput. Tracers are short-lived agents bound to a selector that report back as new connections, profiles and content surface.
The Nerve Centre
Analysts speak only to the DarpanX Nerve Centre, which orchestrates avatar, bot, storage and processing servers behind VPNs, proxies and a DMZ. Offline sources such as mobiles, computers and CDRs enter through the same perimeter. DarpanX was built under the guidance of the National Forensic Sciences University.
