Skip to content
ZeptoDB

Pricing

Community

Free — no license key required

Production-ready for single-node time-series workloads and Agent Memory pilots.

  • Core engine: ingestion, query, ASOF JOIN, LLVM JIT
  • Agent Memory v0: memories, context retrieval, exact/semantic cache
  • Client-supplied embedding storage and ranking
  • Python zero-copy, Python memory/cache helpers, C++ API, Arrow Flight
  • SQL, HTTP API, FIX / ITCH / Binance feed handlers
  • Parquet HDB on S3, GCS, or local NFS
  • Agent Memory sidecar snapshots
  • Basic RBAC, TLS 1.3, API key auth, rate limiting
  • Single-node deployment
  • Community support through GitHub Issues and Discussions

Get started →

Enterprise

License key required — Talk to Sales →

For teams that need clustered time-series, SSO, compliance evidence, deployment support, and a name on the pager.

Everything in Community, plus:

  • Multi-node time-series clustering with automatic sharding
  • SSO / OIDC integration (Okta, Auth0, Azure AD, Google Workspace)
  • Audit log export for SIEM integration
  • Advanced RBAC with per-tenant namespace isolation
  • Kafka connector
  • Migration tools (kdb+ / ClickHouse / DuckDB / TimescaleDB)
  • Geo-replication for time-series deployments
  • Rolling upgrades with zero downtime
  • Production guidance for sticky /api/ai/* routing while Agent Memory remains single-node
  • Priority support with SLA
  • Long-term support releases

Talk to Sales →

30-day free trial

Section titled “30-day free trial”

Try every Enterprise feature for 30 days — no credit card, no sales call required.

Terminal window
# Generate a trial license (single-node, all Enterprise features enabled)
curl -X POST http://localhost:8123/admin/license/trial \
  -H "Authorization: Bearer $ADMIN_KEY"

Trial keys are single-node (max_nodes=1) with all Enterprise features enabled. After 30 days, the instance gracefully downgrades to Community edition.

Ready to run it across a cluster or bring Enterprise into procurement? Contact Sales →

FAQ

Section titled “FAQ”

Is ZeptoDB Community really free? Yes. Community includes the full single-node time-series core, SQL, HTTP API, Python DSL, Arrow Flight, Parquet HDB, and Agent Memory v0. Many teams use it for prototypes, pilots, and single-node production systems.

What requires an Enterprise license? Multi-node time-series clustering, SSO/OIDC, audit log export, advanced RBAC, Kafka connector, migration tools, geo-replication, rolling upgrades, LTS releases, and priority support.

Is Agent Memory cluster-wide today? No. Agent Memory v0 is single-node. In a cluster, route /api/ai/* to a sticky pod or treat it as a per-pod cache. Cluster-consistent memory routing and replicated memory writes are follow-up work.

Does ZeptoDB call LLM or embedding providers? No. Applications supply embeddings and call model providers. ZeptoDB stores memory records, ranks candidates, assembles context, and performs exact/semantic cache lookup.

How does license validation work? Fully offline — RS256 signature verification with no phone-home. Load via environment variable (ZEPTODB_LICENSE_KEY), file (/etc/zeptodb/license.key), or HTTP API (POST /admin/license).

What happens when a license expires? You get a warning before expiry, then a grace period where the deployment continues working. After the grace period, the instance downgrades to Community edition. No data is locked.

Do you offer managed cloud hosting? Not yet. It is on the roadmap. For now, see our Production Deployment and Kubernetes Operations guides, or talk to us about a bespoke deployment.