Pricing

Documentation Index

Fetch the complete documentation index at: https://docs.memorycrystal.ai/llms.txt

Use this file to discover all available pages before exploring further.

Some people need a tiny memory box. Some need a much bigger one. Pricing is how Memory Crystal decides how much memory, how many messages, and how many hosted features each account gets. It now tracks two retention concepts separately: short-term message retention and raw sensory-memory retention.

​

What this means in practice

The product currently has plan lanes such as:

• Free
• Starter / Pro-style paid lanes
• Ultra
• Unlimited / enterprise-style higher-capacity setups

The most important practical differences are usually:

• how many memories can be stored
• how many short-term messages are retained
• how long raw sensory memory content is retained before only its summary/recall text remains
• whether there are channel limits
• how managed Gemini usage is capped or allowed

​

How it actually works

The public product pages talk about pricing at a high level, but the repo-level source of truth for tier behavior is the shared limits model. Important source files:

• shared/tierLimits.ts
• convex/crystal/geminiGuardrail.ts
• relevant backend enforcement paths in convex/crystal/*

That means the docs should describe pricing in public/product language, while still grounding the technical claims in the actual tier limit definitions.

​

Commands / examples

If you want to inspect the technical tier model directly in the repo, start with:

• shared/tierLimits.ts
• README.md

​

Common mistakes

• treating marketing copy as the technical source of truth for limits
• assuming hosted-tier behavior and self-hosted behavior are the same thing
• forgetting that storage limits, message retention, raw sensory retention, and model usage are separate controls

​

Source of truth

Primary files behind this page:

• shared/tierLimits.ts
• README.md
• apps/web/app/pricing/page.tsx
• apps/web/app/data/releases.ts