07 DHARMA · RAKṢĀ

AI Security Scanning That Finds What Others Miss

रक्षा"Protection" — Sanskrit

Semgrep + LLM reasoning in one pipeline. RAKṢĀ combines static analysis with AI-powered contextual understanding to find vulnerabilities traditional scanners can't detect — then generates patches automatically. Now with a full Web UI for drag-and-drop code uploads and one-click GitHub repo scanning.

🛡️SOC 2 Type II

💳PCI-DSS v4.0

📋ISO 27001

🏛️Cyber Trust Mark

🎯MITRE ATT&CK

🔐OWASP Top 10

Start Free Trial See Live Demo ↓

Zero-days found in research

97%

Detection confidence

4-8×

Cheaper than incumbents

<5min

CI/CD integration

Live Demo

Watch RAKṢĀ Scan in Real Time

Click Run to see how RAKṢĀ combines Semgrep static analysis with LLM reasoning to find vulnerabilities and auto-generate patches.

raksha — security scan

Click ▶ Run Scan to see RAKṢĀ in action

Architecture

Three-Stage Intelligence Pipeline

Traditional scanners find patterns. RAKṢĀ understands intent, context, and exploitability — then tells you exactly how to fix it.

RAKṢĀ three-stage security scanning pipeline: static analysis, LLM reasoning, and remediation

DṚṢṬI — Static Analysis

दृष्टि — "Vision"

Semgrep rules scan your entire codebase for known vulnerability patterns, dependency issues, and misconfigurations. Fast, deterministic, zero false negatives on known patterns.

PARIKṢĀ — LLM Reasoning

परीक्षा — "Examination"

DeepSeek R1 and Qwen3 reason about each finding in context — understanding business logic, data flow, and actual exploitability. Reduces false positives by 80%+ compared to rule-only scanners.

YUKTI — Remediation

युक्ति — "Strategy"

Auto-generates code patches, maps findings to MITRE ATT&CK techniques, and produces SARIF reports. Patches are context-aware — they understand your framework and coding style.

Capabilities

Everything Your Security Team Needs

🔍

Code Vulnerability Scanning

Multi-language SAST with Semgrep + LLM hybrid. Supports JavaScript, TypeScript, Python, Go, Java, C/C++, and Solidity.

🎯

MITRE ATT&CK Mapping

Every finding is automatically mapped to ATT&CK v19 tactics and techniques. See which threats your codebase is exposed to.

🌐

Web UI — Scan Anything

Upload ZIP/TAR archives or paste a GitHub URL to scan instantly from your browser. No CLI needed — results appear in real time with severity breakdown and exportable reports.

🔧

Auto-Generated Patches

AI generates context-aware code fixes that understand your framework, dependencies, and coding style. One-click apply.

📊

SARIF Output

Industry-standard SARIF 2.1.0 output integrates directly with GitHub Code Scanning, VS Code, and your existing tools.

📊

Enterprise Security Dashboards

Executive, threat intelligence, and enterprise compliance dashboards with real-time risk scoring, vulnerability trends, and remediation tracking.

🏗️

Infrastructure Posture

Scan Terraform, Kubernetes manifests, Dockerfiles, and cloud configs for misconfigurations and compliance gaps.

📋

Compliance Automation

SOC 2, PCI-DSS, ISO 27001, and HIPAA compliance reporting. Automated evidence collection, audit trails, and regulatory gap analysis.

👁️

Real-Time File Watchers

Watch directories for changes and scan automatically. Get instant security feedback as you code — shift-left security that keeps up with your workflow.

🔒

On-Prem Deployment

Run entirely on your infrastructure with open-source LLMs. Your code never leaves your network. Air-gapped mode available.

Integration

5 Minutes to Your First Scan

Add RAKṢĀ to your CI/CD pipeline with a single GitHub Action. Results appear as native GitHub Code Scanning alerts.

.github/workflows/raksha.yml

name: RAKṢĀ Security Scan

on:
  pull_request:
    branches: [main, develop]
  push:
    branches: [main]

permissions:
  security-events: write
  contents: read

jobs:
  raksha-scan:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - name: Run RAKṢĀ Scan
        uses: avyay-ai/raksha-action@v1
        with:
          api-key: ${{ secrets.RAKSHA_API_KEY }}
          scan-path: ./src
          severity-threshold: medium
          sarif-output: results.sarif
          mitre-mapping: true
          auto-fix: true

      - name: Upload SARIF to GitHub
        uses: github/codeql-action/upload-sarif@v3
        with:
          sarif_file: results.sarif

      - name: Comment PR with findings
        if: github.event_name == 'pull_request'
        uses: avyay-ai/raksha-action/comment@v1
        with:
          sarif_file: results.sarif

GitHubGitLabBitbucketJenkinsCircleCIVS Code

Case Study

How RAKṢĀ Defends Against Real Threats

Operation Cyber GuardianAPT UNC3886

Singapore Telco Defence: All 4 Major Operators Attacked

In February 2026, China-nexus APT group UNC3886 attacked all four of Singapore's major telecommunications operators. Here's how RAKṢĀ's detection pipeline addresses the attack vectors used.

Attack Vector → RAKṢĀ Response

Zero-day exploitation of public-facing applications

T1190 — Exploit Public-Facing Application

RAKṢĀ: PARIKṢĀ LLM analysis detects suspicious input handling patterns before they become zero-days. 97% confidence on SQL injection and command injection vectors.

Rootkit deployment for persistent access

T1014 — Rootkit

RAKṢĀ: DṚṢṬI infrastructure scanning detects kernel module modifications and suspicious system call hooking patterns in IaC configs.

Detection evasion and indicator removal

T1070 — Indicator Removal

RAKṢĀ: YUKTI compliance engine tracks expected vs. actual log integrity. Missing audit trails trigger immediate high-severity alerts.

100+

Cyber defenders mobilised

Months of investigation

4/4

Telcos compromised

ATT&CK techniques used

Source: CSA Singapore, "Operation CYBER GUARDIAN" press release, 9 Feb 2026. RAKṢĀ detection capabilities based on MITRE ATT&CK v19 mapping.

Why RAKṢĀ

Built Different

Feature

RAKṢĀ

Traditional SAST

AI contextual reasoning

✓

✗

Auto-generated patches

✓

✗

MITRE ATT&CK mapping

✓

Partial

SARIF output

✓

On-prem / air-gapped

✓

Rare

Bring your own LLM

✓

✗

False positive reduction

80%+

N/A

Starting price

$2K/mo

$10K+/mo

Pricing

Security That Doesn't Break the Budget

4-8× more affordable than CrowdStrike or Palo Alto. Open-source LLMs + lean architecture = enterprise-grade security at startup prices.

Starter

आरम्भ

$2,000/month

For small teams getting started with AI security

✓Up to 50 repositories
✓Up to 500 endpoints
✓Threat intel dashboard
✓Code vulnerability scanning
✓Basic compliance reports
✓SARIF & JSON output
✓GitHub Action integration
✓Email support

Start Free Trial

Professional

व्यावसायिक

$5,000/month

For growing teams with compliance needs

✓Up to 200 repositories
✓Up to 2,000 endpoints
↑Everything in Starter, plus:
✓Infrastructure posture assessment
✓MITRE ATT&CK mapping
✓Auto-generated remediation patches
✓API access & webhooks
✓Priority support (< 4hr SLA)
✓SOC 2 & PCI-DSS reporting

Start Free Trial

Enterprise

उद्यम

Custom

For large organizations & regulated industries

✓Unlimited repositories & endpoints
↑Everything in Professional, plus:
✓On-prem / air-gapped deployment
✓Bring your own LLM models
✓Custom compliance frameworks
✓Cyber Trust Mark (CTM) audit prep
✓Dedicated success engineer
✓24/7 support (< 1hr SLA)
✓SSO / SAML integration

Contact Sales

Enterprise-Grade Compliance

Built for regulated industries. RAKṢĀ helps you achieve and maintain compliance across major security frameworks.

🛡️

SOC 2 Type II

Compliant

💳

PCI-DSS v4.0

Validated

📋

ISO 27001

Certified

🏛️

Cyber Trust Mark

Level 5 Ready

🎯

MITRE ATT&CK

v19 Mapped

🔐

OWASP Top 10

Full Coverage

Protected by:🛡️ SOC 2🇪🇺 GDPR🔒 CCPA🔐 AES-256Learn more →

Alpha Access · रक्षा

Get Early Access to RAKṢĀ

Join our alpha program. Limited spots — we'll review applications and send API keys to approved users.