Enterprise Features, Security, and Compliance
Built for SOC 2 Type II, ISO 27001, and GDPR from day one. Five-layer isolation protects your data at every level.
Compute
Sandboxed containers
Network
Isolated network segments
Crypto
Dedicated encryption keys (Pro)
Egress
Domain-level traffic filtering
Audit
Comprehensive logging + archival
Prompt Injection Defense
AI assistants that process external content — emails, documents, web pages — are vulnerable to prompt injection attacks, where hidden instructions attempt to hijack the AI assistant's behavior. FixedCostAgents detects and neutralizes these attacks so your AI assistant follows your instructions, not an attacker's.
- Injection attempts hidden in emails, documents, and web content are detected and neutralized before your AI assistant acts on them
- Your AI assistant's identity and instructions are protected — attempts to override its behavior are blocked
- Defenses are continuously updated as new attack techniques emerge in the wild
- Built into the platform — no configuration required, nothing for your team to manage
- Included across all tiers — Standard, Premium, and Pro
- Enterprise customers can request a detailed security brief on our detection and response processes
SOC 2 Type II
Designing for certification
All five Trust Services Criteria (Security, Availability, Processing Integrity, Confidentiality, Privacy) mapped to platform controls. Container sandboxing, envelope encryption, per-customer access scoping, audit logging, and automated provisioning/deprovisioning.
- Sandboxed container runtime with hardened kernel
- Envelope encryption at rest (shared or dedicated key)
- Per-AI assistant access role scoping
- Audit log retention (90-day hot, long-term cold archival)
- Dual-layer ingress authentication (HMAC + bearer token)
- Outbound domain filtering per tier
- Automated provisioning and deprovisioning via billing webhooks
- Hourly data snapshots (RPO 1hr)
ISO 27001
Control mapping complete
Relevant Annex A controls mapped to platform features. Pro tier dedicated infrastructure provides strong control boundaries for network segmentation (A.13) and cryptographic controls (A.10).
- A.5 — Information security policies enforced via infrastructure-as-code
- A.8 — Asset inventory via comprehensive tagging strategy
- A.9 — Access control via RBAC + least-privilege policies
- A.10 — Envelope encryption, dedicated keys for Pro
- A.12 — Automated deployment, continuous monitoring
- A.13 — Network segmentation, subnet isolation, hardware-enforced boundaries
- A.14 — Infrastructure-as-code provisioning, version-controlled configs
- A.15 — Sub-processor inventory (AWS, Stripe, Cloudflare)
GDPR Article 28
DPA template available
Data processor obligations addressed by architecture. Single-region deployment with EU option, customer data ownership, self-service export, and automated 30-day deletion with cryptographic key destruction.
- Data residency: US primary, EU region option
- Encryption at rest and in transit (TLS 1.2+)
- Self-service JSON/ZIP data export via portal
- Automated 30-day deletion on churn
- Cryptographic key destruction (irreversible)
- Per-customer access controls and audit logging
- Sub-processor list with change notification
- 72-hour breach notification commitment
How many AI assistants do you need?
Get a dedicated AI assistant for leaders one level below where leaders have human assistants. Consider one per team or per department. Technical teams may want a group assistant as well as individual ones. They all communicate with each other in your group chats to coordinate and achieve goals together.
Need a security brief, DPA, or vendor questionnaire? Hundreds of AI assistants? Reach out, we’d love to hear from you.
Contact Us