Memory

We differentiate ourselves by layering several kinds of smart retrieval—style, files, and memories—alongside your full account customization. Instead of one bland memory bucket, you get Adaptive Retrieval-Augmented Generation Memory plus Automatic Memory Management: tone and style from your past chats, relevant excerpts from your uploaded files, and the memories that matter for what you just said. Combined with your account settings and preferences, that stack shapes every response. It's how we make the chat feel like it actually knows you.

Memory is how the chat remembers who you are and what you care about across conversations. Without it, every chat would feel like talking to someone who has never met you. With it, the assistant can recall your preferences, work context, and the details you've chosen to store so you don't have to repeat yourself.

What we have today is Adaptive Retrieval-Augmented Generation Memory. Instead of pasting a full memory list into every prompt, the runtime retrieves the most relevant memory chunks for the active scope and injects that smaller memory context into the chat.

Each scope has a Memory toggle. In a project, that main project-memory toggle controls whether project memory participates at all. Automatic Memory Management is the separate background-agent toggle for creating, updating, consolidating, and deleting memories after prompts.

Foreground chat models receive relevant memory context when Memory is on and can use a read-only memory tool for scoped counts, listing, search, and expansion. Memory writes are handled by the background agent, which checks nearby scoped memories and writes only when the change can be grounded in user-authored evidence.

The live system is tier-based and RAG-first: memories can be up to 200 words each, and caps are per scope by plan - 15 on Explorer, 30 on Builder, 75 on Studio, 100 on Aurora, and 200 on Pro. Global, project, and persona layers stay distinct.

Adaptive Retrieval-Augmented Generation Memory is the current memory runtime. The chat embeds recent conversation context, retrieves the most relevant memory chunks for the active scope, and injects that smaller memory context into the prompt instead of stuffing the full memory list every turn.

This keeps memory useful at larger scale. Memories can be up to 200 words each, caps vary by plan, and retrieval counts are configurable in the backend instead of hard-coded in the docs. New accounts are created with this mode enabled.

Global and project scopes stay separate. Persona instructions stay separate too, and persona memories are handled through persona-scoped tools rather than being merged into the user memory buckets.

Automatic Memory Management is the background-agent layer underneath the main Memory toggle. In a project, first turn project memory on. Once that scope is on, Automatic Memory Management can create, update, consolidate, or delete memories for that same scope.

The foreground model still receives relevant memory context through retrieval, plus a read-only memory tool for scoped metadata and inspection. If the user asks to remember, update, or forget something, the assistant can acknowledge naturally while the background agent handles the actual operation.

This keeps memory writes scoped and auditable: project chats use project memories, global chats use global memories, persona writes require an active persona, and memory-change badges let you review what changed.

Automatic Memory Management is the current background memory path. It starts with each eligible prompt, decides whether memory processing should run, and then creates, updates, consolidates, or deletes memories without requiring an in-turn foreground memory tool.

The foreground model receives relevant retrieved memory context when Memory is on and may use a read-only memory tool to inspect scoped memory metadata or contents. It does not receive memory write/edit/delete tools; the background agent handles memory changes.

The current control model is simple: Memory is the master on/off switch, and Automatic Memory Management controls automatic writes. Manual memory edits in the account UI remain available.