Self-Hosted
Deploy ZenSearch on your terms
Free Lite edition for evaluation and small internal use. Custom enterprise deployments for production — every deal is scoped to your environment. Your data never leaves your network.
Lite Edition
Self-host the engine in your VPC. Free for evaluation and small internal use. Up and running in under 10 minutes.
Prerequisites
- Docker Desktop (or Docker Engine + Compose v2)
- 8 GB RAM minimum (16 GB+ recommended)
- Ollama auto-installed — or bring your own LLM (OpenAI, Anthropic, Groq, LM Studio)
What's Included
- Core API, Web UI, Model Gateway
- Document parser + structure analyzer (lite backends)
- Vectorizer + projector pipeline
- S3 + Web Crawler connectors
- PostgreSQL, Redis, Qdrant, RustFS, NATS
- All agent features: multi-agent delegation, three-layer memory, auto-recovery, smart cost routing
- Up to 10 users / 25K documents (soft caps)
- Closed-binary Docker images — EULA, no telemetry, no phone-home
Enterprise self-host
Production deployments — scoped to your environment.
Every enterprise deployment is custom: sized to your user count, compliance posture, and connector mix. The same source binary ships everywhere; enterprise features activate via license key — no redeployment required.
What's available in enterprise deployments
- All connectors (17 total)
- OCR + image vision parsing
- Cross-encoder reranker + sparse embedder
- Semantic paragraph merging + code chunking
- Premium connectors (SAP, Salesforce, MS365 service-account)
- SSO (SAML/OIDC) + SCIM provisioning
- Audit logs + RBAC matrix
- Air-gapped deployment
- Kubernetes + Helm charts with high availability
- Dedicated CSM + technical account manager
- Custom SLA up to 24×7 support
- Compliance review (SOC 2 / HIPAA / GDPR)
- Custom connector development
- Prometheus + Grafana monitoring stack
Talk to our team
Every deployment is shaped to your data, identity stack, and compliance program. Tell us what you need and we'll come back with a concrete proposal.
Contact SalesOr email us directly at sales@zensearch.ai.
Lite vs Enterprise
| Feature | Lite (free evaluation) | Enterprise self-host |
|---|---|---|
| Deployment | Single server (Docker) | Docker / Kubernetes / air-gapped |
| User cap | 10 (soft cap) | Sized to deployment |
| Document cap | 25K (soft cap) | Sized to deployment |
| Air-gapped | Yes (Lite is closed-binary) | Yes |
| Bring your own LLM | Yes | Yes |
| Premium connectors (SAP, Salesforce, MS365) | No | Yes |
| OCR / vision / reranker | No | Yes |
| SSO / SCIM / audit logs | No | Yes |
| Support | Community | 8×5 → 24×7 with dedicated CSM |
Air-Gapped Support
Run ZenSearch with zero internet access. Pre-download container images and model weights. No telemetry, no external calls.
Bring Your Own LLM
Connect Ollama, LM Studio, Groq, OpenRouter, or any OpenAI-compatible endpoint. Use your own GPU infrastructure or hosted gateways for complete model control.
CLI Installer
One-command setup that verifies prerequisites, pulls container images, initializes databases, and starts all services.
Hybrid Mode
Keep data on-premise while using cloud LLM providers for inference. Best of both worlds — data sovereignty with model quality.
Your data. Your infrastructure.
Start with the free Developer Edition, or talk to us about Enterprise deployment.