Tip: Max 10 parallel tasks supported.

Pricing Plans

Choose the perfect plan for your AI memory needs.
Start free and scale as you grow.

Memory Model
Input (per 1K tokens)
Output (per 1K tokens)

Supported Models

gpt-4.1-mini
$0.00040
$0.00160
deepseek-v3.1
$0.00055
$0.00165
gemini-3-flash
$0.00050
$0.00300

Embedding Model

Model
Price (per 1K tokens)
Notes
Voyage 3.5 Lite
$0.00002
Used for embedding search in Memory APIs

Frequently Asked Questions

How does the memory counting work?

Each piece of information stored in MemU counts as one memory. This includes user conversations, preferences, and contextual data.