GalaxDB Cloud·Coming soon.
Join waitlistGalaxDB Cloud
The AI-native database, fully managed. SQL, vector search, local embeddings, and training export with zero infrastructure to run.
Free tier forever. No credit card. Deploy in any region.
// WHY GALAXDB CLOUD
All the power. None of the ops.
Instant setup
One pip install or curl command. No YAML, no cluster config, no cloud console. Your app is running in under a minute.
Local embeddings
No OpenAI API. No Pinecone. Embeddings run inside the database on every INSERT automatically. GDPR-compliant by default.
Time-travel built in
Pin snapshots with CREATE VERSION TAG. Query historical data with AT VERSION. Every training run is fully reproducible.
SQL you already know
Full PostgreSQL wire protocol. Your psycopg2 and SQLAlchemy code works unchanged. Add SEMANTIC_MATCH anywhere in a WHERE clause.
AES-256 at rest
Every PAX block and WAL record encrypted before it hits storage. Pluggable KMS: AWS, GCP, Azure, HashiCorp Vault, or any shell command.
Global regions
Deploy to AWS us-east-1, eu-west-1, ap-southeast-1, and more. Data stays in the region you choose. Multi-region replication on Enterprise.
// GALAXDB CLOUD PRICING
Simple, honest pricing
All cloud plans include the full GalaxDB feature set: SQL, vector search, local embeddings, time-travel, and training export. What scales is capacity and compute, not features.
Free
Start building with the full GalaxDB feature set. No card, no time limit.
- 1 project
- 500 MB storage
- 50K vector ops / month
- 10K embedded rows
- 1 GB training export / month
- Shared compute
- Community support
Pro
More capacity, dedicated support, and no GalaxDB branding.
- 3 projects
- 8 GB storage
- 500K vector ops / month
- Unlimited embedded rows
- 10 GB training export / month
- Shared burstable compute
- Email support
- No GalaxDB branding
- +$0.25 / GB extra storage
- +$0.50 / million extra vector ops
Team
Dedicated compute, SOC 2, and AI-specific billing add-ons.
- 10 projects
- 50 GB storage
- 5M vector ops / month
- 100 GB training export / month
- Dedicated compute (2 vCPU / 8 GB)
- Team management + SSO
- SOC 2 compliance
- Priority support
- Semantic cache: $0.001 / hit
- Near-dedup: $0.50 / million rows
Serverless
Pay only for what you use. Scales to zero between requests.
- $0.00001 / read operation
- $0.00005 / write operation
- $0.10 / 1K embeddings
- $0.023 / GB-month storage
- $0.10 / GB training export
- $0.001 / semantic cache hit
- Auto-scales compute
- No minimum monthly charge
All plans include AES-256 encryption, WAL crash safety, and the complete single-node GalaxDB feature set. No feature gates.
GalaxDB Enterprise
You buy a license and run it on your own infrastructure. For organisations that need distributed scale, compliance certifications, and enterprise governance on top of the Apache 2.0 open-source core.
// BEYOND THE OPEN SOURCE CORE
- Distributed clustering with Raft replication and consistent-hash sharding
- Distributed approximate nearest-neighbor search (correct global top-K)
- Cross-shard consistency and HTAP read replicas
- Automated storage tiering: NVMe, object storage, cold archive
- SSO (OIDC / SAML), fine-grained RBAC, and audit logging
- Federated queries with differential privacy
- Active learning SQL: FEEDBACK, ORDER BY ACTIVE_LEARNING(), drift detection
- Annual commit discounts: 10% to 40%
The open-source single-node engine is free forever. See the full roadmap.
Get in touch
Fill in the details below and we will get back to you within one business day.
// PRICE CALCULATOR
See exactly what you will pay.
Adjust your expected usage below. The estimate updates in real time and shows how it compares to the five-service alternative.
// ADJUST YOUR USAGE
// ESTIMATED MONTHLY COST
vs. the 5-service stack
Estimates only. Actual costs may vary based on region and usage patterns.
// THE MATH
Stop paying for five tools to do one job.
Most AI apps run on PostgreSQL, Pinecone, Redis, S3, and an orchestration tool. That stack costs $370 to $1,500 a month before you have written a single user-facing feature. GalaxDB Cloud Team covers everything those five services do, at $149 a month.
// WHAT YOU STOP PAYING FOR
Three bills you can cancel on day one.
GalaxDB runs embeddings locally, caches semantic queries inside the database, and exports training data with one SQL command. No third-party services needed for any of it.
Save on embedding APIs
$0.10 / 1K embeddings
Embeddings run inside the database using any HuggingFace model. Your data never leaves your infrastructure. No OpenAI API calls, no per-token billing, no latency from a round trip.
Save on LLM query costs
$0.001 / cached query
Repeated semantic queries are answered from a local cache instead of hitting your LLM again. If you send 1M queries a month with a 60% cache hit rate, that is around 600K fewer LLM calls.
Save on data pipelines
$0.10 / GB exported
Training data exports as a versioned Lance dataset with one SQL command. No Airflow, no S3 upload scripts, no custom ETL. Load directly into PyTorch with zero-copy memory mapping.
Be first in line.
Early access opens to waitlist members first. Free tier included, no credit card needed. We will send you a confirmation right away.