{ simple pricing }

Start free. Scale when
your memory grows.

Pay for what your application remembers and recalls. Every plan includes the full API, SDKs, export controls, and no model lock-in.

Start building free
Usage overview
Live

MEMORY RECALLS

184,392

+18.4% this month
Retrieval68% of included usage

Free

For prototypes and personal projects

$0/ month
Start free
  • 10K stored memories
  • 25K recalls / month
  • 1 workspace
  • Community support
  • Full API + SDKs
  • Data export

Scale

For growing production workloads

$249/ month
Start free
  • 2M stored memories
  • 5M recalls / month
  • Unlimited workspaces
  • 90-day event logs
  • Priority support
  • Regional routing

Enterprise

For control, security, and scale

Custom
Contact sales
  • Custom usage
  • Private cloud or on-prem
  • SSO and audit logs
  • Regional deployment
  • Dedicated support
  • DPA + custom terms

USAGE CALCULATOR

Estimate your monthly bill.

Move the sliders to model a workload. This is an illustrative estimate — usage caps and overage are configurable per workspace.

1.2M memories
640K recalls
6 workspaces
Estimated planScale~$249 / month + metered overage

COMPARE PLANS

Feature matrix.

Everything in every plan, side by side. Upgrade, downgrade, or cancel anytime.

FeatureFreeBuilderScaleEnterprise
Stored memories10K250K2MCustom
Recalls / month25K500K5MCustom
Workspaces110UnlimitedUnlimited
Event log retention—7 days90 days1 year+
Per-user auth—✓✓✓
Regional routing——✓✓
SSO / SAML———✓
Audit log export——✓✓
Private / on-prem———✓
SupportCommunityEmailPriorityDedicated

Every plan includes the essentials.

Full REST API TypeScript & Python SDKs MCP server Data export No model lock-in Encryption by default Webhooks Namespaces

PRICING FAQ

Questions teams ask before they sign up.

What counts as a stored memory?

A memory is a single structured fact, preference, event, or relationship extracted from your input. Raw messages are not counted — only the memories they produce.

What counts as a recall?

A recall is one ranked context fetch for a user, namespace, and query. Bulk listing and export calls are not counted as recalls.

Can I bring my own model?

Yes. MatiAI is model-agnostic. Use any provider for extraction and generation — there is no model lock-in on any plan.

Do you offer annual billing?

Yes. Annual plans come with two months free and are available on Builder, Scale, and Enterprise.

Is there a free trial of paid plans?

The Free plan never expires. Paid features can be trialed for 14 days with no card required.

How does overage work?

On paid plans you can cap usage or auto-scale. Overage is metered at the plan's per-unit rate, shown in the dashboard before you incur it.

{ Start building }

Give your AI a memory it can trust.

Free to start. Open by design. Ready for production.

Get MatiAI