Elasticsearch 5–8. OpenSearch 1–3.
Any version. Any direction.
In a weekend.

The open-source migration engine for Elasticsearch and OpenSearch. Raw index-to-index data movement across every version combination. No vendor lock-in. Apache 2.0.

kravex — migration
$ kravex migrate --config kvx.toml

Switching between Elasticsearch and OpenSearch
means migrating all your existing data.

"Just stand up a new cluster and re-index everything."

Whether you're moving from Elasticsearch to OpenSearch or the other way around, you're sitting on 5-50+ TB of production data that must come with you. Compliance mandates retention. Customers depend on historical search. And no existing tool handles this migration at TB-scale without weeks of custom engineering.

PCI DSS 1 year minimum log retention
SOC 2 Retention policies must be enforced and auditable
HIPAA 6-year retention mandate
FedRAMP 1-year minimum, often longer
SOX 7-year retention for financial data

You've already looked at the alternatives.
Here's why they didn't work.

Elasticdump
~1 GB/hour throughput.[1] multielasticdump supports --parallel across indexes, but single-index throughput is limited by the scroll API.[2] No checkpoint/resume. Fine for dev — impractical for production TB-scale.
~1 GB/hr
Logstash
Deep-pagination degradation: 6,000 docs/sec drops to 700 docs/sec on billion-document indexes.[3] Weeks of tuning required to maintain throughput at scale.
88% throughput loss
Reindex API
Cannot migrate mappings, settings, or templates.[4] Source and destination must share direct connectivity (remote whitelist). Versioning and vendor compatibility problems between ES and OpenSearch forks.
ES-only
AWS Migration Assistant
The most capable tool — locked to Amazon OpenSearch Service as the only target. Requires CloudFormation, ECS, MSK, S3. Under 4 TB/day live capture.[5]
AWS-only
Airbyte
Generic ETL. No ES schema optimization, no shard topology handling.[6] Per-row pricing — 10 TB costs thousands in credits.[7]
Generic ETL
AI-generated scripts
Works for 10K docs. Silently loses data at 10M. Crashes at 100M when scroll context times out.[8] No checkpoint/resume, no backpressure, no retry logic.
Silent data loss
Custom in-house tooling
Many organizations have built bespoke ES/OpenSearch migration pipelines. Multi-month engineering projects with ongoing maintenance burden.[9] Kravex replaces this with a single weekend.
3+ months

Two steps. One weekend.

Configuration-as-code. Resume if anything fails.

01 — CONFIGURE

Define your migration

Declare source, target, and runtime settings in a TOML config. Version-controllable, reviewable, repeatable.

# kvx.toml

[runtime]
queue_capacity = 8
sink_parallelism = 4

[source.Elasticsearch]
url = "http://localhost:9200"
index = "employees"

[sink.OpenSearch]
url = "http://localhost:9201"
index = "employees"
02 — MIGRATE

Run the migration

Checkpoint/resume means if anything fails — network blip, node restart, disk full — you pick up exactly where you left off. Not from zero.

$ kravex migrate --config kvx.toml

employees ████████████████ 100%
logs-2024-01 ████████████████ 100%
logs-2024-02 ████████████████ 62%
logs-2024-03 ████████████████ queued

Throughput: ~240 GB/hr | ETA: 6h 12m

One-time purchase. Not a subscription.

Start with the open-source core. Upgrade when you need TB-scale parallelism, compliance features, or white-label embedding.

OPEN SOURCE

Community

Free
Apache 2.0 — forever
  • Full CLI migration engine
  • Config-as-code (TOML)
  • Community support (GitHub, Discord)
Get Started
PRO

Production

Contact
One-time license
  • Everything in Community
  • All connectors and schema mapping
  • Checkpoint/resume on failure
  • Dry-run validation
  • Index-level filtering
  • Basic type mapping
  • Docker image included
  • TB-scale parallel shard processing
  • Intelligent backpressure management
  • Real-time terminal UI dashboard
  • Multi-index orchestration
  • Post-migration integrity checks
  • Priority email support (24h SLA)
Contact Sales
ENTERPRISE

Compliance

Custom
Direct sales
  • Everything in Pro
  • Role-based access control
  • Full audit logging
  • Compliance-ready (SOC 2, HIPAA, PCI)
  • Dedicated SLA support
  • Custom connector development
Contact Sales
Professional services available
PARTNERSHIPS

OEM / White-Label

Custom
Revenue share or flat OEM
  • Everything in Enterprise
  • Full white-label rebrand
  • Embeddable migration widget
  • Partner API (programmatic)
  • Target-specific optimizations
  • Telemetry & reporting dashboard
  • Co-branded documentation
  • Dedicated integration engineering
  • Joint go-to-market support
Become a Partner
Stop losing deals to migration inertia

Your compliance framework requires log retention.
Kravex makes migration possible without breaking the chain.

The only migration tool built for compliance-aware data migration between Elasticsearch and OpenSearch.

PCI DSS

1 year

Minimum log retention. Audit trails must be preserved during and after migration.

SOC 2

Auditable

Retention policies must be enforced, documented, and verifiable by auditors.

HIPAA

6 years

Protected health information retention mandate. No gaps allowed.

FedRAMP

1+ years

Minimum one year, often extended. Continuous monitoring logs included.

SOX

7 years

Financial data and audit logs. The longest standard retention window.

Built for the migrations everyone else ignores.

Every Deployment. Every Direction.

If it speaks the Elasticsearch or OpenSearch API, Kravex can move it. Source or destination — doesn't matter.

Managed services: AWS OpenSearch Service, Elastic Cloud, Aiven, Instaclustr (NetApp), Bonsai, DigitalOcean, OVHcloud, Oracle OCI — any managed offering that exposes an ES or OpenSearch endpoint.

Self-hosted: Kubernetes via ECK or the OpenSearch K8s Operator, Docker Compose, Helm charts, bare-metal RPM/DEB installs, Ansible-provisioned VMs — however you run your clusters.

On-prem: Air-gapped data centers, private cloud, behind a VPN — Kravex is a CLI binary with a TOML config. No SaaS dependency, no phone-home, no cloud account required.

Other tools lock you into a single vendor's ecosystem. AWS Migration Assistant only targets Amazon OpenSearch Service. Elastic's reindex API requires direct cluster connectivity. Kravex works across every combination — managed to self-hosted, on-prem to cloud, or anything in between. Your infrastructure choices are yours. The migration tool shouldn't be the constraint.

Ready to migrate?

Start with the open-source CLI. Upgrade when you need TB-scale performance.