Taskade EVE Memory — Persistent Context Stored as Projects

Overview

Taskade EVE keeps persistent, long-term memory about your apps, agents, and preferences. Unlike most AI assistants, Taskade EVE's memory is not a black box: it lives as a normal Taskade Project inside a folder called projects/memories. You can open it, read it, edit it, or reset it at any time.

TL;DR: Taskade EVE's memory is a real Project in projects/memories. Every fact Taskade EVE remembers is a task you can see, change, or delete. Same Memory → Intelligence → Execution loop as the rest of Workspace DNA.

Why store memory as Projects

• Auditability: Every remembered fact is visible as a task, not hidden in a vector DB.
• Editability: Rename, reorder, or delete tasks and Taskade EVE's next answer uses the new state.
• Portability: Memory travels with the workspace. Share a workspace, share its memory.
• Consistency: The same data powers Taskade EVE, your agents, and your automations.

Where to find it

1. Open the workspace sidebar.
2. Expand the projects folder.
3. Open memories.

Inside you'll see one Project per memory topic. Recent prompts, preferences, and learned facts about your builds live there.

Here is what the memories Project looks like. Each task is one fact Taskade EVE reads before it answers and writes back to after it learns something new:

┌────────────────────────────────────────────────────┐
│ projects  ›  memories                              │
├────────────────────────────────────────────────────┤
│ ▸ preferences                                      │
│     Always deploy to staging first          read   │
│     Keep replies short and friendly         read   │
│ ▸ brand voice                                      │
│     Tone: confident, warm, no jargon        read   │
│     Primary color: Taskade Pink             read   │
│ ▸ decisions                                        │
│     Chose Stripe for invoices               saved  │
│     Renamed "Clients" to "Customers"        saved  │
│                                   [ + Add a fact ] │
└────────────────────────────────────────────────────┘

Taskade EVE reads your preferences and brand voice as it works, and writes new decisions back to this folder so the next build remembers them.

Common memory patterns

Edit memory

Open a memory Project and treat it like any other Taskade Project:

• Add a task to teach Taskade EVE a new fact.
• Delete a task to make Taskade EVE forget.
• Group tasks under subheadings to organize what Taskade EVE remembers.
• Mention the memory Project inside a Taskade EVE prompt with @memories/app-style to force Taskade EVE to read it first.

Reset or rebuild memory

Ask Taskade EVE: "Reset my memory for this app." It moves the current memory tasks to an archived list and starts fresh. Nothing is destroyed; you can restore any past memory by un-archiving it.

See memory inside the Workspace DNA loop

Picture memory as the first turn of a wheel. Memory feeds your agents. Your agents trigger automations. Those automations finish work and write new facts back to memory. That self-reinforcing circle is Workspace DNA: Memory plus Intelligence plus Execution, all running on the same Projects.

Because all three stages share one set of Projects, a fact you edit in memory updates what your agents know and what your automations do. Learn how the full loop fits together in Workspace DNA.

Related guides

• Taskade EVE
• Taskade EVE Mentions
• Taskade EVE Slash Commands
• Agent Knowledge
• Workspace DNA: why Memory, Intelligence, and Execution live in one loop

Deep reading

• Memory Reanimation Protocol: why AI agents forget and how Taskade fixes it
• Workspace DNA Architecture: the self-reinforcing Memory to Intelligence to Execution loop
• The 2026 Productivity Playbook: the hub for agents, automations, and Taskade Genesis apps