Frequently Asked Questions
Everything you need to know about MythosLab
Privacy & Data Control
Does MythosLab store any of my data?
No. Absolutely not. MythosLab is 100% self-hosted. Everything runs on your hardware—the agents, the database, the dashboard, all of it. We never see your data, we never store your data, and we never transmit your data to our servers. There are no analytics, no phone-home mechanisms, and no cloud dependencies.
What exactly stays on my machine?
Everything:
- All AI agents and their inference models
- Your conversations, task history, and preferences
- RAG knowledge bases (Athena, Themis agents)
- Secrets and credentials (Hecate vault)
- Business documents, contracts, and analytics data
- The web dashboard and API
- The embedded SQLite database
MythosLab is designed to run entirely on your business infrastructure. The only external connections are the ones you explicitly configure (e.g., connecting to your email, Slack, or business tools).
Can I use cloud-based LLMs?
Yes, but it's optional. MythosLab is designed to work primarily with local inference via Ollama (85-90% of tasks). For complex reasoning tasks, you can optionally configure API keys for cloud providers like Claude, Gemini, ChatGPT, or Grok. When you do this, only the specific prompts are sent to the provider—your historical data, preferences, and context remain local.
Remote Access for Your Team
How do I access MythosLab remotely?
MythosLab runs locally at localhost:4340. To access it remotely, you set up your own secure tunnel. We recommend three approaches:
Zero-config VPN. Install on your server and devices. Access via private network. No ports to open.
Creates a secure tunnel without exposing your IP. Free tier available.
Full control with self-managed VPN and reverse proxy. For experienced users.
📖 Detailed setup guide: We'll provide comprehensive documentation on setting up each option. For now, see Tailscale's official guide for the quickest path.
Does MythosLab provide a hosted dashboard?
No. The dashboard runs on your machine as part of the Docker container. This ensures your data never leaves your control. Remote access is your responsibility—we provide the tools, you choose how to expose them (or not).
Technical Requirements
What are the system requirements?
Minimum (Basic agents):
- 4 GB RAM
- 20 GB disk space
- 2 CPU cores
- Docker or Podman
Recommended (Local inference with Ollama):
- 16 GB RAM (32 GB for large models)
- 50 GB disk space
- 6+ CPU cores or NVIDIA GPU
- Linux, macOS, or Windows with WSL2
Can I run this on a small office server?
Absolutely. MythosLab is designed for self-hosted deployment on modest hardware. A standard office server or even a capable desktop PC will handle the full business agent suite with local inference. See the system requirements above for specifics.
What tech stack does MythosLab use?
MythosLab follows the SnowyStack specification:
- Language: Python
- Framework: Django
- Database: SQLite
- Interactivity: HTMX
- Styling: Tailwind CSS
- Inference: Ollama (local) + cloud LLM APIs (optional)
Features and Integrations
How do I upload legal documents for Themis analysis?
Navigate to Knowledge → Legal Documents in the header menu to access the dedicated legal document repository. Themis (our legal & compliance agent) processes these documents in a separate RAG collection for compliance isolation.
Supported use cases include:
- Contract review and clause extraction
- Regulatory compliance analysis
- Policy document comparison
- Legal precedent search
Supported formats: PDF, TXT, MD, CSV, HTML (max 50MB per file).
How do I configure CI/CD webhooks for Hephaestus?
Navigate to Admin → Webhooks to access the webhook administration panel. Hephaestus (our DevOps agent) supports webhooks from:
- GitHub: Push, pull request, and release events
- GitLab: Push, merge request, and pipeline events
- Jenkins: Build started, completed, and failed events
The webhook panel displays the full URL endpoints for each provider. Copy the URL, add it to your repository settings, select the events to monitor, and save. Test events will appear in the recent events table.
Hephaestus can trigger automated build analysis, test failure notifications, and deployment workflows based on webhook events.
General Questions
When will MythosLab be available?
MythosLab is currently in active development. Join our waitlist to be notified when beta access opens. We're targeting Q2 or Q3 for initial release, and we'll update this as we get closer. (If you're on the waitlist, you'll be amongst the first to know... ).
Will MythosLab be open source?
We're evaluating licensing options. Our goal is to make MythosLab accessible to privacy-conscious users while sustaining development. We'll announce our decision before the beta launch.
How much does it cost?
MythosLab offers three subscription tiers designed for small businesses:
- Core: $49/month — 3 core agents (Zeus, Athena, Hermes), RAG
- Business: $99-299/month — All 17 agents, business automation, analytics, compliance
- Home: $49-99/month — Core + home automation, personal productivity, smart home
With tiered inference, your total AI compute cost is typically about $15-45/month on top of the subscription — versus $600-1,800/month for a 20-seat team on per-seat enterprise AI. See how we calculate this. Beta testers will receive early-adopter discounts.
Can I extend MythosLab with custom agents?
Yes! MythosLab is designed to be extensible. You'll be able to create custom agents using the same Python/Django architecture. Documentation for the agent SDK will be available at launch.
Still have questions?
Join our waitlist and we'll keep you updated as we get closer to launch.