AI Memory Comparison

Interactive comparison of persistent memory systems for AI tools. Two categories, two different sets of concerns:

  • Developer Tools — Memory for coding assistants (Claude Code, Cursor, etc). MCP support, search approaches, storage backends, dedup, cross-machine sync, token budgets.
  • User Memory Platforms — Memory for agents and chatbots (Mem0, Zep, Letta, Cognee). Pricing, SDKs, multi-tenancy, temporal graphs.

How it works

Each memory system is a JSON file in the data/ directory:

data/
├── developer-tools/
│   ├── husk.json
│   ├── claude-mem.json
│   └── ...
└── user-memory/
    ├── mem0.json
    ├── zep.json
    └── ...

Files are validated at build time with Zod schemas defined in src/lib/types.ts. The site renders them into interactive TanStack Tables with sorting, filtering, column toggling, and shareable URL state.

Contributing

Add a new system

  1. Create a JSON file in data/developer-tools/ or data/user-memory/
  2. Follow the schema — look at any existing file as a template
  3. Open a PR

The build will validate your JSON against the Zod schema automatically.

Update existing data

Edit the relevant JSON file and submit a PR. Include a source link for any claims.

Schemas

Each JSON file includes a $schema reference for editor validation. JSON Schemas are generated from the Zod source of truth at build time (pnpm generate-schemas):

Zod definitions: src/lib/types.ts

Development

pnpm install
pnpm dev

Open http://localhost:3000.

License

MIT