AI Chatbots
- Assistants with access to your knowledge base
- Connect to files, wikis, databases, and internal tools
- Create personas per team or use case with policy controls
- Prompts, history, and embeddings remain encrypted
Making Private AI Practical
Run AI on sensitive data—without exposing it. Enclava executes your prompts, tools, and RAG pipelines inside confidential computing enclaves so data stays encrypted in use, at rest, and in transit.
Open source • Privacy‑first architecture • On‑prem or private cloud
Deploy Enclava open-source on your own infrastructure
Join waitlist for our fully managed cloud platform
Discover Enclava's privacy-focused AI features
What you can build
Real workloads teams run on Enclava today—secure by default, developer‑friendly by design.
Document Processing
- Ingest PDFs, Office docs, emails, and text at scale
- Ask questions and get cited answers with source links
- Organize into secure collections with role‑based access
- Sub‑second semantic search across your private corpus
Plugins & Integrations
- Build secure tools with a lightweight SDK
- Key budgeting, rate limits, and granular access per workspace
- Drop‑in connectors for Slack, Jira, GitHub, Notion, and more
Confidential Computing
- Run models and RAG inside trusted execution environments (TEEs)
- Memory and storage are encrypted by default
- Remote attestation to verify code and environment
- Operational model where even we can't access your data
Workflows
- Orchestrate multi‑step agents, tools, and retrieval
- Parallel execution, retries, and idempotency built‑in
- Transform data between steps with typed adapters
- Schedule runs or trigger via webhooks
Developer APIs
- OpenAI‑compatible endpoints for chat, RAG, and embeddings
- REST APIs for everything, with simple auth
- Usage, cost, and performance visibility per project
- Extend the platform with custom plugins
A great fit when you have…
PII, PHI, or financial data
Handle sensitive personal information with confidence
Strict privacy requirements
Meet regulatory standards without compromising on functionality
Compliance & audit obligations
Built-in compliance features for enterprise governance
Proprietary information you want to keep confidential
Protect your intellectual property and trade secrets
Why Enclava
01
Privacy‑first architecture
End-to-end encryption in use, at rest, and in transit. Your data never leaves your control, even during AI processing.
02
Verifiable security
Remote attestation and cryptographic proof that your code runs in genuine secure enclaves. Trust, but verify.
03
Drop‑in compatibility
OpenAI‑compatible APIs and familiar developer tools. Migrate your existing AI workflows in minutes, not months.
04
Enterprise‑ready
Built-in user management, audit logs, usage controls, and enterprise SSO. Scale from prototype to production.
Frequently Asked Questions About Private AI
Learn more about confidential computing, secure AI deployment, and how Enclava protects your sensitive data.
What is confidential computing for AI?
Confidential computing uses hardware-based Trusted Execution Environments (TEEs) like Intel SGX and AMD SEV to protect AI models and data during processing. This ensures complete privacy even from cloud providers, system administrators, and potential attackers.
How does Enclava ensure HIPAA compliance for healthcare AI?
Enclava provides end-to-end encryption, comprehensive audit logging, role-based access controls, and runs all processing in isolated TEEs. This meets all HIPAA technical safeguards for protected health information (PHI), enabling secure AI processing of medical records, patient data, and clinical research.
Can I self-host Enclava on my own infrastructure?
Yes, Enclava is open source and can be self-hosted on your infrastructure. We provide Docker images, Kubernetes manifests, Terraform modules, and comprehensive deployment guides for on-premise, private cloud, or hybrid deployments.
What's the difference between Enclava and VPN encryption?
VPNs only encrypt data in transit between endpoints. Enclava uses TEEs to encrypt data during processing (in use), in addition to encryption at rest and in transit, providing complete protection throughout the entire AI pipeline including model inference and RAG operations.
Which AI models and LLMs does Enclava support?
Enclava supports variety of open source models - gpt-oss-128b, llama 3.3, gemma3 and qwen2.5 coder
How does secure RAG (Retrieval-Augmented Generation) work in Enclava?
Enclava's secure RAG pipelines process documents, embeddings, and vector searches either on your own infrastructure (self hosted) or within TEEs(managed cloud). Your knowledge base remains encrypted, and retrieval operations happen in isolated enclaves, ensuring that sensitive documents are never exposed during augmentation.
Is Enclava suitable for financial services and PCI DSS compliance?
Yes, Enclava's architecture supports PCI DSS requirements with encrypted processing, access controls, audit logging, and network segmentation. Financial institutions can process payment data, customer information, and transaction records with AI while maintaining compliance.
What's the performance impact of using confidential computing?
Modern TEE implementations like Intel TDX and AMD SEV-SNP have minimal overhead (typically 5-15%) for most AI workloads. Enclava optimizes enclave operations and supports GPU acceleration where available, ensuring practical performance for production use cases.
Can Enclava integrate with existing vector databases?
Yes, Enclava uses popular vector database Qdrant. It supports externally managed collections so you can use your own RAG pipeline with Enclava.
How do I get started with Enclava?
You can start with our open-source self-hosted version on GitHub, join the waitlist for our managed cloud service, or book a consultation for custom deployment. We offer comprehensive documentation, deployment guides, and professional services to help you implement private AI.
Join the Enclava Cloud waitlist
Get early access to our fully managed cloud platform. No infrastructure setup, no maintenance—just secure private AI ready to use.
Cloud platform only. For self-hosted options, see deployment section below.
Get started with Enclava
Choose the deployment option that works best for your team and security requirements.