Pricing
One data layer, unmetered SQL, unlimited seats, and always-on AI. Choose flexible usage or reserved capacity—without a per-seat tax.
On-Demand
14-day free trialPay only for what you use. Best for flexible or growing workloads.
Up to 5TB Compressed Storage
Up to 512 Compute Units
Unlimited SQL Queries
Best-effort Compute Cache
APIs & MCP Access
Unlimited Seats
Email Support
Reserved
Volume & commitmentReserve capacity upfront. Predictable pricing, performance, and volume discounts.
Custom Storage
Custom Compute Units
Unlimited SQL Queries
Bring Your Own Compute
Warm Compute Cache
Up to 365-day Time Travel
APIs & MCP Access
Unlimited Seats
VIP Support
Estimate your monthly cost.
Bare-metal Performance. Unlimited SQL Queries. Always-on AI.
Compressed Storage
High-performance object storage & NVMe cache.
100GB
~200.0M rows
Compute Units
Compute capacity with maximum performance.
32
~115,200 medium-complexity queries per 24h
AI Units
SOTA token capacity for hosted AI workloads.
20M tokens
20 AI Units
Compare plans in detail
Same product surface area; on-demand and reserved differ mainly by capacity, cache behavior, support, and commercial terms.
Swipe sideways to compare On-Demand and Reserved.
| Feature | On-Demand | Reserved |
|---|---|---|
| Lakehouse | ||
| Compressed storage | Up to 5 TB | Custom |
| Compute units | Up to 512 ACUs | Custom |
| SQL queries | Unlimited | Unlimited |
| Compute cache | Best-effort caching | Pre-warmed cache |
| Bring Your Own Compute New | - | |
| Time travel New | - | Up to 365 days |
| Volume discount | - | |
| AI | ||
| SOTA LLM | Sonnet 4.6, Opus 4.7, GPT 5.5 | Custom |
| Semantic Layer | ||
| Claude Plugin | ||
| MCP | ||
| Product Analytics | ||
Event Tracking API /track | ||
User Identification API /identify & /alias | ||
| Segmentation, Funnel & Retention Analysis | ||
| SQL Access | ||
| Custom SDKs (TypeScript, Swift, Kotlin, Ruby, Python, etc.) | ||
| Data Analytics | ||
| Chart, Table, and Metric Visualizations | ||
| Dashboards | ||
| Custom SDKs (Python, Ruby, TypeScript, Java, etc.) | ||
| External Connections | ||
| PostgreSQL | ||
| MySQL | ||
| Supabase | ||
| MariaDB | ||
| AWS S3 Bucket | ||
| Cloudflare R2 Bucket | ||
| Google Cloud Storage Bucket | ||
| Apache Iceberg | ||
| BigQuery | ||
| Snowflake | ||
| Integrations | ||
| DBT | ||
| Hex | ||
| Metabase | ||
| Looker | ||
| Tableau | ||
| Custom (API, SDK, etc.) | ||
| Platform | ||
| HTTP API | ||
| Role-based access control | ||
| Seats | Unlimited | Unlimited |
| SSO | Coming soon | Coming soon |
| Support | Email support | VIP support |
Frequently asked questions
How does Altertable pricing work?
Altertable pricing is based on the three resources you use: storage, compute, and AI. You pay for compressed data, ACUs, and AI Units for model usage.
What is included in every plan?
Every plan includes unlimited seats, API and MCP access, dashboarding, product analytics ingestion, and external data connections. Plans mainly differ by capacity, support, and commercial terms.
How is storage measured?
Storage is billed on compressed data stored in Altertable, not raw source size. Analytical data often compresses to 10-30% of its original size.
What is an Altertable Compute Unit?
An Altertable Compute Unit (ACU) is a share of analytical compute capacity for queries, transformations, and background work.
What is Bring Your Own Compute?
Bring Your Own Compute (BYOC) allows you to run Altertable's compute engine on your own infrastructure while still benefiting from Altertable's hosted control plane.
Can I connect my own data sources to Altertable?
Yes. You can connect your own databases, warehouses, S3 Buckets, Iceberg Tables, and more to Altertable as external catalogs.
Do you charge per seat?
No. Seats are unlimited, so pricing scales with platform usage rather than headcount.
What is the difference between On-Demand and Reserved compute?
On-demand compute is billed by actual usage. Reserved compute gives you committed baseline capacity for more predictable pricing and steadier performance.
Why not bill per query like a warehouse?
Altertable is priced for continuous intelligence workloads, not one-off queries. You pay for available compute capacity rather than each question or dashboard refresh.
How is AI priced?
AI is priced in AI Units. Each AI Unit equals 1M blended input, cached, and output tokens.
A lakehouse your apps, BI, and agents share
DuckDB workers on open formats, federated SQL across your existing systems,
and an MCP server for agents — at flat monthly pricing.