Simple pricing that scales with you
Start free. Pay as you grow. Predictable pricing with transparent overage rates on every plan.
Free
For individual developers exploring AI memory.
- 1 project
- 3K memories
- 6K recalls / month
- 1K ingests / month
- 2K stores / month
- 1 API key
- Auto hibernate after 2 days
- Email support (72 hrs)
Starter
For small teams shipping agents to production.
- 5 projects
- 50K memories
- 80K recalls / month, then $0.002/op
- 10K ingests / month, then $0.005/op
- 30K stores / month, then $0.0005/op
- 5 API keys
- 7-day backup retention
- Auto hibernate after 7 days
- Email support (48 hrs)
Pro
For teams running memory at scale with full cognitive features.
- 20 projects
- Unlimited memories
- 1M recalls / month, then $0.001/op
- 100K ingests / month, then $0.004/op
- 300K stores / month, then $0.0004/op
- Unlimited API keys
- 30-day backup retention
- Auto hibernate after 14 days
- Reflect / Consolidate: 1K/mo, then $0.01/op
- Email support (24 hrs)
Enterprise
For organizations with compliance, scale, and deployment requirements.
- Unlimited instances
- Custom instance size (up to 16 GB)
- Volume operation discounts
- Dedicated nodes (optional)
- VPC peering available
- 99.9% uptime SLA
- 90-day + custom backups
- Dedicated CSM
- SSO / SAML included
- Audit logs included
Frequently asked questions
A memory is a single unit of stored information: a conversation turn, an extracted fact, an insight, or a summary. The cognitive pipeline may generate multiple memories from a single ingest call.
On Starter and Pro plans, you can continue using the service at the listed overage rates. On the Free plan, operations are paused until the next billing cycle. You never lose data.
Inactive projects are automatically hibernated after the specified period to save resources. They wake up instantly on the next API call. No data is lost during hibernation.
Memory data is encrypted at rest (AES-256) and in transit (TLS 1.3). We never use customer data for training. Enterprise plans include additional controls like SSO/SAML and audit logs.
No. DeltaMemory uses proprietary models for extraction, retrieval, and cognitive processing. Everything is handled for you — no API keys to manage, no model selection, no prompt engineering.
Reflect generates insights from recent memories. Consolidate merges similar memories into summaries. Both are cognitive operations available on the Pro plan and above.
Yes. You can upgrade or downgrade at any time. When upgrading, you get immediate access to the new plan. When downgrading, the change takes effect at the end of your current billing cycle.