InventDB

Start free. Grow on a plan sized to you.

Request the free tier of InventDB Serverless or InventDB SOAR here. Paid plans are priced to your requirements: tell us what you need and we send you a price.

Granted on approval

Start free, no card.

Request the free tier of either product, we review it, and confirmation arrives by email.

The database

InventDB Serverless Free

100,000 records · 1 GB storage · 1 GB RAM · full SQL, search, files and MCP included · sleeps when idle.

Free on approval Request free InventDB Serverless
The platform

InventDB SOAR Free

The same limits, plus a $10 monthly AI credit. Workflows and scheduled reports are not limited.

Free on approval Request free InventDB SOAR

One free instance per account, with the same encryption and permissions as a paid plan. Moving to a paid plan is a plan change and does not involve any migration.

Every paid plan

The fine print.

Whole machine: every instance is single-tenant on AWS Graviton; its memory and processors are the whole machine and not a share of one. Storage is your own encrypted data volume, AES-256-GCM under your own key; the record count of your plan is a per-instance maximum. System disk is a second volume holding the OS and a swapfile sized at twice the instance RAM, so a query that outgrows memory slows down instead of being killed. Transfer is a monthly allowance per plan; go over it and transfer throttles for the rest of the cycle instead of adding to your bill. AI allocation on InventDB SOAR is fixed per plan; when it is used up the AI pauses for the rest of the month, so you are never billed past your plan. On every plan: encryption, the audit trail, region pinning, no training on your data, and full export. Not offered: a HIPAA BAA or on-prem deployment.

Choosing between the two.

InventDB Serverless

The database alone, on your own encrypted volume: full SQL, ACID, search and vectors. It has no AI inside, and external agents connect over MCP.

InventDB SOAR

The same database with the business engine and console on top. AI is included on every plan, and every answer shows the SQL behind it.

The same engine

Both run one engine per instance and scale from a few thousand rows to 250 million records with no migration between plans.

Best for

Developers who want a database to build on choose InventDB Serverless; teams who want to run a business without building an app choose InventDB SOAR.

FAQ

Common questions.

How is a paid plan priced?

Per instance, per month, sized to what you need: records and storage, memory and processors, data transfer and, on InventDB SOAR, the AI allowance. Write to contact@inventdb.com and we send you a price.

How does billing run?

One subscription is one instance, billed monthly in advance; dev, staging and prod each subscribe separately. Past the transfer allowance, transfer throttles, and on InventDB SOAR the AI pauses when the month's allowance is used, so nothing is added to your bill.

Can I change plans later?

Yes, in either direction. Plans are arranged with us: write to contact@inventdb.com, and we resize your instance in place and confirm the new price. Switching products means starting a new subscription.

What happens if I cancel?

Write to contact@inventdb.com to cancel. You get a full export of your records, files and saved templates. We delete your database and key 30 days after cancellation and confirm it in writing.

Does InventDB Serverless include AI?

No. It is a database with no AI inside, and your data is never sent to any AI model. Every instance ships a built-in MCP server so you can point any AI client at it; workflows, report authoring, the integrations and plain-English analytics live on InventDB SOAR.

Which region runs my instance?

Pinned by your signup country: US customers run in a US AWS region, EU customers in an EU region, India and everywhere else in AWS Mumbai. Data stays in-region on your own encrypted volume, and on InventDB SOAR the Claude AI runs on Amazon Bedrock in that same region.

Is my data used to train models?

No. On InventDB SOAR the AI runs with the no-training opt-out through the InventDB AI Gateway. Your prompts and data never enter any training pipeline.