Enterprise
April 21, 2026 · View on GitHub
Pricing anchors at $3,000/month and scales with load, seats, and compliance requirements. For teams who need private deployment, regulatory controls, and dedicated support.
Contact: enterprise@divergencerouter.com
Who This Is For
- Regulated industries — healthcare (HIPAA), finance (SOC2 Type II, PCI), government (FedRAMP-adjacent)
- High-volume production — 10M+ chunks stored, 1K+ eval runs per month
- Privacy-sensitive — customer data that cannot leave your VPC under any circumstance
- Multi-tenant platforms — you're embedding agent memory into your own product
- Teams > 25 engineers — when the Business tier's 10-seat cap stops fitting
What Enterprise Includes
Everything in Business, plus:
Deployment
- VPC / private cloud deployment — observability + eval backend runs inside your AWS/GCP/Azure environment
- On-premise option — air-gapped environments supported for government and defense
- Kubernetes manifests + Terraform — infrastructure-as-code, no ClickOps
- BYO-LLM — bring your own Ollama/vLLM cluster for consolidation, or use our managed workers
Identity + access
- SSO / SAML 2.0 — Okta, Auth0, Azure AD, Google Workspace
- SCIM 2.0 provisioning — automate user lifecycle from your IdP
- Role-based access — admin, workspace owner, editor, viewer, auditor roles out of the box
- Custom role definitions — fine-grained permission grants
Compliance
- Signed DPA (Data Processing Addendum) — standard + customizable terms
- Signed BAA (Business Associate Agreement) — HIPAA workloads
- SOC2 Type II — report available under NDA; Type I on Business tier
- PCI DSS compliance — for credential and financial type chunks
- GDPR / CCPA — deletion SLA, data residency options
- Audit logs — every chunk add/delete/search, with configurable retention (default 2 years)
Security
- Customer-managed encryption keys (CMEK) — you hold the keys; rotate on your schedule
- Private network only — no public internet egress
- IP allowlisting — API access restricted to your ranges
- Hardware security module (HSM) support — for credential-type chunks
- Penetration testing reports — annual 3rd-party report available
Support + SLA
- 99.95% uptime SLA with service credits
- 24/7 dedicated support via shared Slack + on-call phone
- Named technical account manager (TAM)
- Quarterly architecture reviews — we help you optimize memory patterns as you scale
- Priority feature requests — your roadmap asks get weight
Custom work
- Custom retention policies — per namespace, per chunk type, per user
- Custom eval scenarios — we build AMB extensions that match your domain
- Custom consolidation prompts — tuned for your industry vocabulary
- Migration assistance — from LangChain / Mem0 / MemGPT / custom systems
- Training — engineering team workshops on memory-driven agent design
Pricing Model
Base fee: starts at $3,000/mo. Includes 10 seats, 1M chunks stored, 500 eval runs/mo, 500 hosted consolidation jobs/mo.
Scales with:
- Seats (adds at $30/seat/mo after 10)
- Storage (adds at $0.20 per 1K chunks/mo after 1M)
- Hosted consolidation jobs (adds at $0.75/job after 500)
- Dedicated support tier (standard, premium, white-glove)
- Regional residency (single-region standard, multi-region +20%)
Typical deals land in the $5K–$50K/mo range depending on scale and compliance.
Annual contracts get a 10% discount. Multi-year (2+ years) get 15%.
Security Posture
- SOC2 Type II (available under NDA)
- Annual 3rd-party penetration test (SANS-certified)
- Bug bounty program (private, scoped)
- Dependency scanning on every build (Snyk + Dependabot)
- Secrets scanning on every commit
- No training on customer data (contractual, not just policy)
- 100% reproducible builds
Reference Architecture
┌──────────────────────────────────────────┐
│ Your VPC │
│ │
│ ┌─────────────────────────────────┐ │
│ │ archon-memory-core OSS library │ │
│ │ (in your application runtime) │ │
│ └──────────────┬──────────────────┘ │
│ │ │
│ ┌──────────────▼──────────────────┐ │
│ │ Control plane (our code, │ │
│ │ your VPC, helm chart) │ │
│ │ - Dashboard API │ │
│ │ - Eval runner │ │
│ │ - Replay debugger │ │
│ │ - Retention engine │ │
│ └──────────────┬──────────────────┘ │
│ │ │
│ ┌──────────────▼──────────────────┐ │
│ │ ChromaDB / your vector store │ │
│ │ (never leaves your VPC) │ │
│ └──────────────┬──────────────────┘ │
│ │ │
│ ┌──────────────▼──────────────────┐ │
│ │ Ollama / vLLM cluster │ │
│ │ (consolidation, your GPUs) │ │
│ └─────────────────────────────────┘ │
└──────────────────────────────────────────┘
│ (opt-in telemetry only)
▼
Our observability cloud
(anonymized metrics, no memory content)
Data flow guarantee: raw memory content never traverses our cloud. We see only anonymized metrics (chunk counts, eval scores, job counts). You turn even those off with a single config flag.
What We Don't Do
- We don't sell your data. Period.
- We don't train models on your memory. Period.
- We don't require exclusivity. You can run our OSS alongside other memory systems. If you leave, your data stays where it is — we never held it.
- We don't lock you in. The OSS library is Apache 2.0 and self-sufficient. If Enterprise stops working for you, your production never stops — you just lose the observability layer.
Procurement
- Legal review: we accept most vendor security questionnaires. Common answers pre-packaged.
- Invoicing: NET 30 standard, NET 60 on multi-year. Procurement portals (Ariba, Coupa) supported.
- POC / pilot: 30-day full-feature pilot with signed NDA. Converts to contract at end, or walks away.
- Tax-exempt: nonprofits and public-sector entities qualify with documentation.
How to Buy
- Email enterprise@divergencerouter.com with a brief description of your use case.
- We schedule a 30-minute scoping call — no slideshow, no sales theater.
- You get a quote + redlined DPA within 5 business days.
- 30-day pilot → annual contract.
For everyone else, start with Business or Team. If you outgrow it, we're here.