Best AI Memory for Windsurf (2026)
Windsurf's Cascade Memories are great inside one workspace and forget across them. Six ways to give Windsurf persistent memory — including the Cascade Memories baseline, the .windsurfrules baseline, and a one-config-block hosted layer — evaluated honestly.
The quick answer
If you want persistent memory that drops into Windsurf in 60 seconds and gives you a browsable wiki of what it remembered: Ricord. If you only ever work in one workspace and like curating: Cascade Memories or a .windsurfrulesfile. If you're building agents on top of Windsurf as a coding environment: Letta. The matrix below spells out the rest.
Why this matters for Windsurf specifically
Cascade is one of the strongest in-editor agents shipping today, and Cascade Memories are a real step up from nothing — Windsurf auto-captures facts and lets you pin your own, scoped to the workspace. The break shows up at the seams: across workspaces, across the other AI tools you use, and over time. A memory you formed in one project doesn't follow you to the next, and nothing you teach Windsurf is reachable from Claude Desktop, Cursor, or ChatGPT.
The fixes cluster into six approaches that differ on a handful of axes — install effort, workspace portability, auto-extraction, wiki browsability, cross-tool reach, conflict resolution, and cost. The right answer is a function of how you use Windsurf day-to-day, not a single winner.
The decision matrix
Eight criteria, six options. The two Windsurf built-ins (Cascade Memories and the .windsurfrules file) are evaluated separately because they answer different problems.
| Criterion | Ricord | Mem0 | Letta | Supermemory | Cascade Memories | .windsurfrules |
|---|---|---|---|---|---|---|
| MCP server install (one config block) | Limited | |||||
| Survives across separate workspaces/repos | Per-workspace | Per-workspace | ||||
| Auto-extracts from your chats (no curation) | Manual | |||||
| Browsable wiki of what was learned | Memories list | Raw markdown | ||||
| Knowledge graph + backlinks | Pro only | |||||
| Same memory in Claude Desktop / Cursor / ChatGPT | API only | |||||
| Conflict resolution (old facts superseded) | Basic | Manual | Manual edits | |||
| Cost (smallest paid tier with memory) | $12/mo annual | $249/mo for graph | Self-host + LLM | $29/mo | $0 (built in) | $0 |
Slot-by-slot — which fits you
If you live in Windsurf daily across multiple workspaces
Ricordis built for this — install once, every Cascade session writes to the same memory + wiki, scoped per project automatically. Windsurf's MCP support picks it up without per-workspace setup. The wiki view is the payoff: by week two you can read what your AI has learned about each codebase, in one place, instead of hunting through per-workspace memory lists.
If you work in one workspace and like the built-in
Cascade Memoriesare genuinely fine here. They auto-capture and you can pin your own, all without leaving Windsurf. The ceiling is portability and browsability: the memories live in that workspace, there's no graph across them, and you can't reach them from your other AI tools.
If you want static project rules, not memory
A .windsurfrules fileat your workspace root is the right tool for stable conventions — style, do-not-touch lists, deploy steps. You write the rules, Cascade follows them, no infra. It doesn't auto-learn from your conversations, but for small static contexts that's a feature, not a bug.
If you're building an agent product on top of Windsurf
Lettaships an agent runtime with memory built in. If you're already writing agent orchestration code, the integrated model is cleaner than wiring memory as a separate service. If you just want memory for Windsurf itself, Letta is overkill.
If you have engineering time to tune retrieval
Mem0open-source is the right pick. Apache-licensed, vector-first, forks cleanly. You'll spend real time getting it production-ready, but if retrieval quality isyour product's edge, owning the code is the right trade. When OSS wins →
If Windsurf is one of many AI surfaces you use
This is the strongest case for a hosted layer over the built-in. Ricord (or Supermemory) keeps one memory that every tool reads — what you teach Windsurf shows up in Claude Desktop and Cursor too. Cascade Memories, by design, stay inside Windsurf.
Why Ricord wins for most Windsurf users
- One install covers every workspace.Memory follows you across projects, scoped automatically so they don't leak into each other — no per-workspace memory list to curate.
- The wiki view is browsable. Open the dashboard and read what Windsurf learned about each codebase — an actual document tree and knowledge graph, not a flat list of memory snippets.
- It reaches your other tools.The same memory is live in Claude Desktop, Cursor, and Codex via Ricord's MCP server. Teach it once in Windsurf, recall it anywhere.
bun add -g ricord ricord login ricord install # auto-detects Windsurf, Cursor, Claude Code, Claude Desktop
Restart Windsurf. Open Cascade. Ask it to remember something. Ask again tomorrow in a different workspace. The wiki populates as you work.
Getting started
Pick the option that matches your slot. If it's Ricord, the three commands above get you running. If it's Cascade Memories or .windsurfrules, you're already set up — just start pinning. The signal that you picked right is whether you've stopped re-explaining the same context to Windsurf after a week.