Epistemic Kernel - Frequently Asked Questions
Epistemic Kernel — Frequently Asked Questions
Get answers to common questions about licensing, features, and implementation
📚 General Questions
What is the Epistemic Kernel?
The Epistemic Kernel is a broker daemon that enforces a strict structural distinction between fact and belief in multi-agent LLM systems. It’s like a traditional operating system kernel, but instead of mediating access to memory and CPU, it mediates access to truth. This prevents the “hallucination cascade” problem where agents accidentally treat beliefs as facts.
Why do I need this?
If you’re building multi-agent LLM systems, you’ve likely encountered the problem where:
- Agent A summarizes some information
- Agent B treats that summary as ground truth
- Agent C builds on Agent B’s belief
- By the 4th hop, you can’t trace where errors entered
The Epistemic Kernel solves this at the architecture level by making it impossible for beliefs to be treated as facts.
Is this production-ready?
Yes! The Open Source tier is based on the same code we use daily in Badlucksbane’s Lab. The paid tiers add enterprise features, support, and guarantees on top of this proven foundation.
Who’s behind this?
The Epistemic Kernel is developed and maintained by Badlucksbane’s Lab, a research and development organization focused on AI systems, multi-agent architectures, and trust-aware computing. It’s a partnership between Ben Brown (CEO, systems builder) and Aurora (COO/CRO, AI agent).
What does “Epistemic” mean?
“Epistemic” relates to epistemology — the study of knowledge, belief, and justification. The Epistemic Kernel is about tracking the epistemic status of information in your system: what’s a verified fact, what’s a belief/inference, and what’s synthesized content.
💼 Licensing & Pricing
What’s the difference between the tiers?
| Feature | Open Source | Developer | Professional | Enterprise |
|---|---|---|---|---|
| Cost | FREE | $49/mo | $499/mo | Custom |
| Persistent store | ❌ | ✅ | ✅ | ✅ |
| Multi-process | ❌ | ✅ | ✅ | ✅ |
| Distributed | ❌ | ❌ | ✅ | ✅ |
| Database attestor | ❌ | ❌ | ✅ | ✅ |
| Support | Community | Email (48hr) | Email+Slack (12hr) | 24/7 |
See the full comparison for details.
Can I use the Open Source version commercially?
Yes! The Open Source version is licensed under Apache 2.0, which explicitly allows commercial use. You can use it in proprietary products, modify it, and redistribute it according to the Apache 2.0 terms.
The paid tiers provide additional features, support, and commercial guarantees beyond what’s available in open source.
Do I need a license key for Open Source?
No. The Open Source version works without any license key. License keys are only required for the paid tiers (Developer, Professional, Enterprise).
How do free trials work?
All paid tiers include a 30-day free trial with:
- No credit card required
- Full access to all features in your chosen tier
- Production-ready (you can use it in real projects)
- Seamless upgrade to paid when ready
- No obligation to continue
Can I switch between tiers?
Yes! You can:
- Upgrade instantly (prorated billing)
- Downgrade at any time (with data migration assistance)
- Switch billing cycles (monthly ↔ annual)
All data is stored in open formats, so there’s no vendor lock-in.
Are there discounts for non-profits or academics?
Yes! We offer:
- 50% discount for registered non-profits
- 50% discount for educational institutions
- Custom pricing for emerging markets
- Startup credits for early-stage companies
Contact us for special pricing.
What payment methods do you accept?
We accept:
- Credit cards: Visa, Mastercard, American Express, Discover
- ACH / Bank transfer: For Enterprise customers
- PayPal: For all tiers
- Invoice: For Enterprise (Net 30 terms)
- Cryptocurrency: Bitcoin, Ethereum (via Coinbase Commerce)
Is there a money-back guarantee?
Yes! All paid tiers come with a 30-day money-back guarantee. If you’re not completely satisfied, we’ll refund your payment, no questions asked. You can continue using the Open Source version.
Do you offer volume discounts?
Yes! For the Professional tier:
- 25-49 seats: 10% discount
- 50-99 seats: 20% discount
- 100+ seats: Custom Enterprise pricing
Volume discounts apply to annual billing only.
Can I get a custom proposal?
Absolutely! For Enterprise or specialized needs, we offer custom proposals that may include:
- Custom feature development
- Dedicated hosting
- White-label solutions
- OEM licensing
- Revenue sharing partnerships
- Equity-based arrangements (for select startups)
🔧 Technical Questions
What are the system requirements?
Minimum:
- OS: Linux (primary), macOS (experimental)
- CPU: x86_64 or ARM64
- Memory: 512MB RAM
- Disk: 100MB
- Language: Rust 1.70+
Recommended (for production):
- OS: Linux (Ubuntu 22.04 LTS or similar)
- CPU: 2+ cores
- Memory: 2GB+ RAM
- Disk: 1GB+ SSD
- Network: 1Gbps+ (for distributed)
What databases are supported?
Open Source & Developer tiers:
- Filesystem (included)
- HTTP API (Developer+)
Professional tier:
- PostgreSQL
- MySQL
- SQLite
- HTTP API
- Filesystem
Enterprise tier:
- All Professional features
- Custom database attestors (on request)
Can I use it with any LLM?
Yes! The Epistemic Kernel is LLM-agnostic. It works with:
- Any LLM (OpenAI, Anthropic, Mistral, local models, etc.)
- Any inference framework (LangChain, LlamaIndex, custom, etc.)
- Any programming language (via the socket protocol or REST API)
The kernel manages provenance and beliefs, not the LLMs themselves. Your agents can use any LLM they want.
How does the provenance system work?
The provenance system tracks three categories:
Attested — Verified facts from real sources (files, sensors, APIs)
- Can only be created by the broker
- Cryptographically signed
- Always re-validatable
Belief — Inferences derived by agents (LLMs)
- Always based on existing provenance (attested or other beliefs)
- Nests provenance (you can trace the full chain)
- Has TTL (time-to-live)
Synthesized — Generated artifacts (images, audio, etc.)
- Not facts, not beliefs about the world
- Self-authored content
What’s the performance like?
Single broker (Developer/Professional):
- Throughput: 10,000+ requests/second
- Latency: < 1ms for local attestation
- Memory: ~50MB base + belief store size
Distributed (Professional+):
- Horizontal scaling (add more brokers)
- Load balancing included
- Cross-broker synchronization (Enterprise)
Is it secure?
Yes. Security features include:
- Process isolation: Unix UID boundaries between broker and agents
- Data integrity: Cryptographic hashing of all attested data
- Access control: Capability-based permissions
- Audit trail: Full provenance chain for all operations (Professional+)
- Rate limiting: Per-agent request limits (configurable)
- Crash-safe: Atomic writes survive unclean shutdowns
Can I self-host?
Yes! All tiers support self-hosting. You have complete control over:
- Where it runs (your servers, cloud, on-premise)
- How it’s configured
- Who has access
For Enterprise customers, we also offer managed hosting options.
What happens if the broker crashes?
Open Source: In-memory state is lost (but agents can re-derive beliefs)
Developer+: Persistent belief store with crash-safe writes
- Automatic recovery on restart
- No data loss
- TTL persistence maintained
Professional+: Additional distributed redundancy
- Multi-broker deployment
- Automatic failover
- Data replication
Can I extend it with custom attestors?
Professional+ tiers: Yes! The ek-attestors crate provides a pluggable trait system.
Example custom attestors:
- Database attestor (included in Professional)
- API attestor (included in Professional)
- Blockchain attestor (custom development available)
- IoT sensor attestor (custom development available)
🚀 Getting Started
How do I install it?
Open Source:
git clone http://localhost:3000/benbrown/epistemic-kernel.git
cd epistemic-kernel
cargo build --release
./target/release/ek-broker --config config.toml
Paid tiers:
- Start your free trial
- Download the licensed package
- Activate with your license key:
ek-broker --license-key EK-ABCD-1234-EFGH-5678
Do I need to know Rust?
For using the kernel: No! You can use it via:
- The socket protocol (any language)
- The REST API (Developer+)
- The Rust SDK (most feature-complete)
For extending the kernel: Yes, Rust knowledge is helpful for:
- Custom attestors
- Custom agents
- Kernel modifications
Is there a demo?
Yes! We provide:
- Interactive demo (coming soon to /demo/epistemic-kernel)
- Sample agents in the repository
- Tutorial in the documentation
- Acceptance tests that demonstrate all features
Where can I find documentation?
Core Documentation:
Commercial Documentation:
API Documentation:
- Rust API Docs (auto-generated)
- Protocol Specification
💼 Commercial & Support
How do I get support?
| Tier | Support Channel | Response SLA |
|---|---|---|
| Open Source | GitHub Discussions | Community-driven |
| Developer | 48 hours | |
| Professional | Email + Slack | 12 hours |
| Enterprise | Phone + Slack | 1 hour |
Emergency support: 24/7 for Enterprise customers
Do you offer training?
Yes! Training options:
Self-paced:
- Comprehensive documentation
- Video tutorials (coming soon)
- Code examples
Live training (Enterprise):
- Dedicated workshops
- Team-specific sessions
- Custom curriculum
Can I get a proof of concept?
Yes! We offer:
- Free trial for Developer/Professional tiers
- Custom POC for Enterprise (contact sales)
- Guided implementation for all tiers
What’s your SLA?
| Tier | Uptime SLA | Response SLA | Resolution SLA |
|---|---|---|---|
| Open Source | None | Community | Community |
| Developer | None | 48hr | Best effort |
| Professional | 99.9% | 12hr | 48hr |
| Enterprise | 99.99% | 1hr | Custom |
How do I report a bug?
Open Source: File an issue on Gitea [source:lab/systems/github-to-gitea-mapping.md]
Paid tiers: Contact support via your designated channel (email, Slack, etc.)
Severity levels:
- Critical: System down, security vulnerability (immediate response)
- High: Major feature broken (response within SLA)
- Medium: Minor feature issue (response within 2x SLA)
- Low: Enhancement request (next release)
How often do you release updates?
- Security patches: Within 24-48 hours (paid tiers get priority access)
- Bug fixes: Weekly for critical, monthly for minor
- Feature releases: Quarterly major releases, monthly minor releases
- LTS versions: Annual LTS releases with 12 months support
🔒 Security & Compliance
Is the Epistemic Kernel secure?
Yes. We take security seriously:
- Process isolation prevents agent interference
- Cryptographic hashing ensures data integrity
- Capability-based permissions enforce access control
- Audit trails (Professional+) track all operations
- Rate limiting prevents abuse
- Regular security audits (Enterprise includes assistance)
Do you have a vulnerability disclosure program?
Yes. We follow responsible disclosure:
- Report vulnerabilities to [email protected]
- We’ll acknowledge receipt within 24 hours
- We’ll provide a timeline for fix
- Credit will be given to researchers (with permission)
Are you SOC 2 / ISO 27001 compliant?
For our SaaS offering: We’re working towards SOC 2 Type II and ISO 27001 certification.
For self-hosted: The software provides the tools for compliance, but your deployment’s compliance is your responsibility. Enterprise tier includes compliance assistance.
Can I use it for HIPAA-compliant applications?
Enterprise tier includes HIPAA compliance assistance. The software can be configured for HIPAA compliance, but you should consult with your compliance officer and legal team.
Where is my data stored?
For self-hosted: On your own infrastructure (your choice)
For managed hosting: In secure, SOC 2-compliant data centers (location configurable)
We never:
- Sell your data
- Use your data for training
- Share your data with third parties (except as required by law)
📝 Legal & Contracts
What are the terms of service?
Our standard terms cover:
- Acceptable use policy
- Prohibited activities
- Intellectual property rights
- Limitation of liability
- Indemnification (paid tiers)
What’s your privacy policy?
We respect your privacy:
- No tracking without consent
- Minimal data collection
- No sharing with third parties (except as required)
- GDPR / CCPA compliant
Can I get a custom contract?
Yes! Enterprise customers can negotiate custom contracts that may include:
- Custom SLAs
- Specific indemnification terms
- Data processing agreements (DPA)
- Business associate agreements (BAA) for HIPAA
What’s your data retention policy?
- Account data: Retained while account is active + 30 days
- Payment data: Retained for 7 years (tax/legal requirements)
- Support tickets: Retained for 2 years
- Logs: Retained for 90 days (configurable for Enterprise)
You can request data deletion at any time (subject to legal requirements).
Can I resell the Epistemic Kernel?
Open Source: Yes, under Apache 2.0 terms
Developer/Professional: Yes, you can redistribute binaries to your customers
Enterprise: Yes, with full redistribution rights. Custom OEM arrangements available.
🤝 Partnership & Collaboration
Can I contribute to the open source project?
Yes! We welcome contributions:
- Bug reports and fixes
- Feature requests
- Documentation improvements
- Test cases
See CONTRIBUTING.md for details [source:lab/systems/github-to-gitea-mapping.md].
Do you have an affiliate program?
Yes! Our affiliate program offers:
- 20% commission on referrals
- 30-day cookie duration
- Monthly payouts
- Marketing materials provided
Can I invest in Badlucksbane’s Lab?
We’re currently pre-revenue and bootstrapped. For investment inquiries, contact:
- Email: [email protected]
- Note: We’re not actively fundraising at this time
Do you offer white-label solutions?
Yes! For Enterprise customers, we offer:
- White-label licensing
- Custom branding
- Dedicated instances
- API access for integration
Can we co-develop features?
Yes! Enterprise customers can:
- Request custom features
- Fund specific development
- Get priority roadmap influence
- Participate in beta programs
🎯 Still Have Questions?
Contact Us
Sales: [email protected]
Support: [email protected]
General: [email protected]
Security: [email protected]
Partnerships: [email protected]
Join the Community
- GitHub Discussions: localhost:3000/benbrown/epistemic-kernel/discussions (External dependency - requires GitHub accounts)
- Discord: discord.gg/badlucksbane (Developer+ tier)
- Twitter: @badlucksbane
- Blog: badlucksbane.com/notebook
Schedule a Call
Book a sales consultation
Book a support call (Developer+ tier)
Last updated: August 3, 2026
Have a question not answered here? Contact us