National Cyber Warfare Foundation (NCWF)login

IntelHub for browser-based OSINT investigations with local AI analysis


0 user ratings
2026-11-17 01:30:31
milo
Red Team (CNA)
"IntelHub

IntelHub is a multi-browser extension that turns your browser into an authorized OSINT investigation suite, combining Telegram profiling, metadata forensics, and a privacy-first local AI agent.








Tooltomsec8/IntelHub — a browser-based OSINT suite with Telegram analysis, digital forensics, and local AI integration, currently at v5.0.2
CategoryOSINT browser extension / reconnaissance toolkit (JavaScript, MIT licensed)
Primary UseAuthorized investigators and analysts using Telegram profiling, EXIF extraction, reverse image search, and a local LM Studio AI agent for documented research on public data
Safe UseLegitimate use is limited to authorized OSINT investigations, defensive research, due diligence, and lab analysis of publicly available information; all features process public data or analyst-supplied files
Telemetry NoteThe extension claims no analytics or telemetry; defenders should note it requests scripting, tabs, and clipboardRead permissions, and that agent-driven searches hit DuckDuckGo HTML, Unshorten.me, corsproxy.io, and VirusTotal — observable as outbound traffic from the analyst workstation

IntelHub from tomsec8/IntelHub is a JavaScript-based browser extension, MIT licensed and sitting at roughly 121 stars, that positions itself as a full OSINT working environment rather than a single-purpose utility. Version 5.0.2 is distributed through the official Chrome Web Store, Firefox Add-ons, and the Microsoft Edge Add-ons store, which matters for an investigation tool because store distribution implies at least cursory review and gives analysts a verified installation path instead of sideloading. The README describes the target audience explicitly: researchers, analysts, and investigators conducting authorized open-source intelligence work.


What distinguishes this release is the Local OSINT Agent. IntelHub v5.0.2 integrates with LM Studio, so an analyst can run a local LLM entirely offline and interact with it through a chat interface in the extension sidebar. The privacy pitch is consistent: prompts, uploaded files, and analysis never leave the local machine because inference happens on the analyst's own hardware. For teams handling sensitive investigations where cloud LLM terms of service are a non-starter, this is the architecture detail worth noting.


The agent is not isolated from the web, however. It performs live search via the DuckDuckGo HTML endpoint, which the README highlights as requiring no API keys — a pragmatic choice, though it means analyst queries do traverse a third-party search provider even when the model itself is local. The tension between offline inference and online enrichment is inherent to the design, and operators should understand which half of the pipeline touches the network. Cross-origin features are gated behind an interactive consent checkpoint according to the privacy policy, which is a reasonable control for that boundary.


The document forensics capability is fully client-side. Dropping a PDF, image, or DOCX into the chat triggers local SHA-256 hashing and metadata extraction, with the AI formatting a structured report from the results. Under the hood this leans on well-known open-source libraries: ExifReader for image metadata, pdf-lib for PDF analysis, jszip for Office document handling, and compromise NLP for entity extraction from text. This is a sensible composition of vetted components rather than bespoke parsers, which reduces the surface for parsing bugs that have historically plagued forensic tooling.


The Telegram Intelligence module covers username-based profiling — details, profile pictures, and bios — plus numeric ID extraction for users, groups, and channels, with the README noting that web login is required for the ID features. There is also a phone lookup that generates quick links to check number registrations. Everything here operates on public or login-visible data, which keeps it firmly in the OSINT lane rather than anything resembling unauthorized access, but analysts should still document their authority for any profiling activity since platform terms of service apply.


Site and archive reconnaissance is the other major cluster. A Digital Fingerprint feature identifies tech stacks, cookies, and user-agent data on the current page; a Safety Checks flow chains URL unshortening through Unshorten.me with optional VirusTotal scanning; and Time Travel queries historical copies across Wayback Machine and Archive.today. The SingleFile integration enables saving a complete local HTML snapshot of a page — a genuinely useful evidentiary function, since snapshots preserve context that link rot destroys.


Text and social analysis round out the toolkit. The Text Profiler extracts emails, crypto addresses, and phone numbers from arbitrary text blocks using compromise, while the Social ID Extractor pulls numeric identifiers from major platforms in one click — this is what consumes the scripting and tabs permissions, injecting into the active tab to read IDs. The Investigation Graph, rendered with Cytoscape.js and Chart.js, lets analysts map entities and relationships visually, which converts scattered lookups into something closer to a structured case file.


From a defender's perspective, the permission set deserves a read before installation: storage, scripting, tabs, clipboardRead, downloads, and activeTab. Every one has a stated purpose in the README, but scripting plus tabs is the classic powerful combination, so enterprise teams should verify the store listing matches the source. The privacy policy also claims that network sniffing tools are domain-scoped and self-terminating, and that external tool definitions fetched from GitHub are schema-validated over HTTPS to mitigate XSS — both statements that a cautious operator would verify against the code in modules/.


The codebase layout is transparent and readable: popup.html and popup.js drive the interface, content.js handles in-page logic, and modules/ separates concerns into telegramAnalyzer.js, siteAnalyzer.js, investigationGraph.js, localAgent.js, and others. Vendored libraries live in libs/. This modularity makes the extension auditable, which is exactly what you want in a tool that touches clipboard contents and active tabs during sensitive work.


The favorites and customization system should not be overlooked as an operational feature. Analysts can pin frequently used tools into custom categories, register their own external tools, and export/import configurations for team sharing — a workflow detail that turns the extension into a shared team baseline rather than a personal utility. Multilingual guides in help/ cover English, Hebrew, Spanish, French, German, Portuguese, and Polish.


Installation is either through the browser stores or by cloning the repository at https://github.com/tomsec8/IntelHub and loading the unpacked extension in developer mode for review. For authorized investigators, defensive researchers, and due-diligence teams that want OSINT conveniences without shipping case data to a cloud AI provider, IntelHub's local-first architecture and auditable module design make it a credible entry in the browser-extension recon space — provided operators respect platform terms of service and the consent boundaries the tool itself tries to enforce.



Official project repository for tomsec8/IntelHub.

Download Tool

Educational analysis for authorized security professionals. Use only in controlled, authorized environments.






Source: OffensiveSec
Source Link: https://www.offsecblog.com/2026/11/intelhub-for-browser-based-osint.html


Comments
new comment
Nobody has commented yet. Will you be the first?
 
Forum
Red Team (CNA)



Copyright 2012 through 2026 - National Cyber Warfare Foundation - All rights reserved worldwide.