Pricing

It's cheaper than you think

A lifetime of journaling fits in under $5 a month. And β€” using Memoria actually drops your LLM input-token bill.

🎁 Beta-only β€” keep a card on file, use the full service for free.

The prices below apply starting from the official launch. During beta, a card on file unlocks every Pro-plan feature with no caps β€” storage, recall, priority, all unlimited. We give at least 30 days of notice before the beta ends, at which point you pick a plan or cancel.

Free

Start with no commitment

$0 /mo

  • βœ“ 100MB storage
  • βœ“ 50 recalls/day
  • βœ“ Standard priority
Get started
Most popular

Personal

Enough for most people, forever

$4.99 /mo

  • βœ“ 5GB storage (β‰ˆ 200 books)
  • βœ“ Unlimited recall
  • βœ“ Unused credit rolls over
Get started

Pro

For heavy users / all-day coding & research

$14.99 /mo

  • βœ“ Unlimited storage
  • βœ“ Unlimited recall + priority
  • βœ“ Bulk backfill / import
Get started

Plan limits and names are finalized just before launch. We email at least 30 days before beta ends.

πŸ’‘ Cost inversion β€” your LLM bill drops as you use it more

LLMs re-send the entire conversation history on every message. The longer the chat, the more input tokens accumulate every turn. With Memoria on, your context is just the new message plus the top 5–10 relevant memories β€” keeping per-turn input around 2K–3K tokens, roughly constant.

Usage pattern Without Memoria (mo) With Memoria (mo) Savings
Light user (20 turns/day) $3–5 $1–2 ~$2
Typical user (50 turns/day, growing context) $15–25 $3–5 ~$15+
Heavy user (coding/research, 100+ turns/day) $80–150 $10–20 ~$100+

* Estimates based on Claude Sonnet 4.6 input pricing at $3 / 1M tokens. Will be replaced with measured values at launch.

A Memoria subscription ($4.99) pays for itself many times over in LLM input-token savings. The heavier the user, the bigger the gap β€” Memoria is infrastructure that cuts cost, not adds to it.

πŸ’Ž What you don't use doesn't disappear β€” Rollover

A monthly subscription where unused capacity gets wasted is a losing structure. We deduct real usage (storage, embeddings, recall calls) and roll the remainder forward as token credit.

Item Amount
Credit added this month $4.99
Embeddings + storage used $1.20
Recall calls used $0.80
Actual usage this month $2.00
Remaining credit β†’ next month $2.99

Accumulated credit can be applied to: (a) next month's overage, (b) priority queue / bulk import / ChatGPT-Claude conversation backfill and similar add-ons, (c) future family / team seat sharing.

FAQ

Who sees my data?
Only you. A self-hosted option is on the roadmap.
ChatGPT, Claude, Gemini β€” does it work with all of them?
Any MCP-capable assistant works directly. Others can be captured via hooks / browser extension.
I'm worried about privacy if everything is captured automatically.
Auto-capture is opt-in, and you can pause or delete by category / session.