ABOUT SYDSEC

Built by a Practitioner,
for Practitioners.

SydSec is a UK-based cybersecurity software company with a simple belief: the best AI security tools should work without an internet connection, without cloud subscriptions, and without sending your sensitive data somewhere you can't control.

We build offline-first AI assistants for red teamers, blue teamers, forensic analysts, and security researchers. People who operate in environments where cloud AI simply isn't an option.

SydSec
Cybersecurity AI, United Kingdom

9.27
Avg Accuracy
14B
LLM Parameters
100%
Offline

The Story

Syd didn't start as a product. It started as a problem.

During penetration tests and security assessments, the same question kept coming up: why isn't there a tool that actually understands what I'm looking at? Not a generic chatbot that hallucinates CVEs. Not a cloud API that uploads your client's network data to a third-party server. Something local, accurate, and built around the tools practitioners actually use.

So we built one. Syd started as a personal project to see whether a locally-running large language model, combined with purpose-built knowledge bases and a deterministic fact extraction layer, could actually help during a real engagement, not just in a demo.

The answer was yes. After months of testing against real Nmap scans, Volatility memory dumps, BloodHound graphs, YARA rules, PCAP files, and NetExec output, Syd v3 hit a 9.27/10 average accuracy score across all production tools. That's when it became a product.

Today, SydSec ships Syd as a self-contained Windows application. No cloud, no subscription, no data leaving your machine. You get a 14-billion-parameter AI assistant that understands your tooling, interprets your output, and gives you answers grounded in curated, tool-specific knowledge rather than whatever was scraped from the internet.

Why Offline-First?

Cloud AI is powerful, but it creates a problem in security work that nobody talks about openly: your queries contain sensitive information.

When you paste a client's network scan into ChatGPT, or upload a memory dump to an online analysis service, that data travels to a third-party server. It may be logged, retained, or used for model training. For many engagements (government, defence, finance, legal) that's simply not acceptable.

Even for smaller engagements, there's a professional obligation to protect client data. Syd is built around this from the ground up. Every component runs locally: the LLM, the vector search index, the fact extractor, the knowledge base. Nothing leaves your machine.

This also means Syd works in air-gapped environments. Isolated lab networks, secure assessment workstations, systems that can never reach the internet by design. Most AI tools simply cannot operate there. Syd was built for exactly that constraint.

Our Approach

Tool-Specific Knowledge

Syd doesn't use a generic model with generic knowledge. Every supported tool (Nmap, BloodHound, Volatility, YARA, Metasploit, Sliver) has its own dedicated knowledge base, curated specifically for that tool's output formats, terminology, and common use cases.

Anti-Hallucination by Design

Hallucination is the biggest risk in AI-assisted security work. Syd uses a three-layer approach: deterministic regex-based fact extraction, constrained LLM prompting, and a validation layer that cross-checks the model's output against extracted facts before showing you any results.

Practitioner-Led Development

Every feature in Syd reflects a real workflow need. The analysis outputs, the knowledge base structure, the GUI layout: all shaped by what's actually useful during an engagement, not what looks impressive in a demo. We test against real tool output, not sanitised examples.

Who We Serve

Syd is designed for security professionals who live inside terminals and command-line tools, not people who manage them from a dashboard.

Penetration Testers
Faster scan analysis, automated CVE correlation, attack path reasoning across Nmap, BloodHound, NetExec, and Metasploit output.
Digital Forensic Analysts
Memory dump analysis via Volatility, YARA rule application, PCAP investigation, all with AI-assisted interpretation of findings.
Blue Team & SOC Analysts
Detection engineering support with Zeek, Suricata, Sysmon, Chainsaw, and TShark, with MITRE ATT&CK alignment built in.
Security Educators & Students
The Community edition is a powerful learning companion. Ask questions, understand tool output, and build expertise without needing cloud access.

UK-Based, Independently Built

SydSec is an independent UK cybersecurity company. We're not a VC-funded startup chasing growth metrics, and we're not a consultancy with a product bolted on. We're a small team building serious tools for serious practitioners, and we're proud of that.

Being independent means we answer to our users, not investors. Features get built because they're useful, not because they're marketable. Pricing is fair because we set it ourselves. When something doesn't work the way it should, we fix it, because our reputation depends on it.

If you have feedback, a feature request, or you're running into something unexpected, reach out directly. We read every message.

Responsible Use

Authorised Use Only

Syd is designed for authorised penetration testing, security research, and incident response. It must only be used against systems you own or have explicit written permission to test.

Unauthorised use against systems you do not own or have permission to access is illegal under the Computer Misuse Act (UK), the CFAA (USA), and equivalent legislation in other jurisdictions. SydSec accepts no liability for misuse.

By purchasing or downloading Syd, you confirm that you have explicit authorisation for any systems you test, you will comply with all applicable laws and regulations, and you will not use Syd to conduct unauthorised access or attacks.

For licensing queries, contact info@sydsec.co.uk

Ready to Try Syd?

Whether you're evaluating Syd for your organisation, have a technical question, or just want to check whether it'll run on your hardware, get in touch. We're happy to talk.