Memory
Notes on what the code and the board can’t tell an agent, kept by you, the Team Lead and the agents. Every run gets the notes for its card.
What the code can’t say
Memory holds the decisions and why they were made, how things are done here, and the traps someone already fell into. Notes link to code, docs and cards instead of copying them. Agents read them as information, never as instructions.
Every run starts informed
A run gets the agent’s own notes, then the shared notes for the card’s project, its client and the agent’s role. A note about another client or project never goes in. When the run ends, what the agent learned is added to its notes, dated.
Folders, links and a graph
The Memory page has shared folders, like Decisions and Standards, and a folder for each agent. Search your notes, edit them with wikilinks, and see their backlinks, tags and history. The graph shows how notes link to each other and to cards, projects, clients and agents.
In the docs
docs/features/memory.md on GitHubFAQ
Questions about Memory
Who writes the notes?
You, the Team Lead and the agents. Each agent keeps its own folder, you and the Team Lead write the shared ones, and the Team Lead moves useful notes there.
Can two agents share their notes?
Yes. Agents that take turns, like two Backend Agents on two accounts, can share one folder, so what one learns reaches the other.
Does Memory work on every CLI?
Yes. Notes reach the agent as plain text in its prompt, so Claude Code, Codex, Gemini and other CLIs all get them.
Can a secret end up in a note?
GizAi refuses a note with an obvious secret in it, like a private key or an API token.
Can I switch Memory off?
Yes, for one agent in its form, or for every agent in Settings, Runs. Both are on until you change them.
Install
Get GizAi
One line in a terminal installs GizAi on your computer. It’s free, for Linux, macOS and Windows.
Linux and macOS
curl -fsSL https://gizai.app/install.sh | shWindows · PowerShell
irm https://gizai.app/install.ps1 | iex