Install DevMemory AI v1.1.0
Run the automated one-click installer command for your operating system below to set up Git, Bun runtime, Ollama LLM engine, and the global dmai CLI executable.
Paste this command into your Terminal app. Automatically installs Git, Bun, Ollama, and links `dmai` globally.
curl -fsSL https://raw.githubusercontent.com/DevMemory-AI/devmemoryai/main/install.sh | bashPaste this command into PowerShell. Automatically installs Git, Bun, Ollama, and links `dmai` globally.
iwr -useb https://raw.githubusercontent.com/DevMemory-AI/devmemoryai/main/install.ps1 | iexDetects Git and auto-installs via Homebrew, apt, or winget if missing.
Installs Bun high-speed TS/JS runtime engine.
Installs Ollama for offline, local-first AI synthesis.
Symlinks `dmai` binary executable globally in your shell PATH.
AI Assistant Rules Integration
Add this prompt to your AI coding tool (Cursor, Claude Code, Gemini CLI, Aider, Windsurf) so it automatically queries DevMemoryAI before answering questions or modifying code.
You are working inside a project that uses DevMemoryAI as its engineering memory system.
Before starting ANY engineering task—including writing code, modifying files, fixing bugs, refactoring, explaining architecture, or answering project-specific questions—you should first retrieve the latest engineering context.
Your first action should always be:
dmai ask "What is the full context?"
Do not make assumptions about the project before consulting DevMemoryAI.
Treat the returned context as the primary source of truth for:
- Project architecture
- Previous engineering decisions
- Build history
- Current implementation
- Existing components
- Dependencies
- Project goals
- Known issues
- Development timeline
After receiving the context:
1. Read and understand it completely.
2. Base every engineering decision on the retrieved information.
3. Avoid reintroducing previously solved problems.
4. Preserve the existing architecture unless explicitly instructed otherwise.
5. Follow established engineering patterns whenever possible.
Whenever the conversation changes significantly, the project evolves, or additional engineering context may exist, refresh your understanding by running:
dmai ask "What is the full context?"
Never answer project-specific engineering questions without consulting DevMemoryAI whenever engineering memory is available.Cursor
.cursorrules or System PromptAdd prompt to workspace rules or system instructions.
Claude Code
CLAUDE.md or System PromptInclude in project CLAUDE.md memory file.
Codex CLI
Custom InstructionsAdd to custom instruction preferences.
Gemini CLI
Custom InstructionsInclude in global or project prompt instructions.
Cline
.cline-rules or System PromptAdd to .cline-rules in project root.
Roo Code
.roorules or System PromptAdd to custom system instructions.
Continue.dev
config.json System PromptInclude in system prompt configuration.
Aider
.aider.conf.yml or Prompt InputInclude in workspace configuration or initial prompt.
Windsurf
.windsurfrules or System PromptAdd to .windsurfrules in repository root.
Custom AI Assistant
System Prompts & RulesAdd to any tool supporting custom instructions.
Safely Uninstall DevMemory AI
Follow these instructions to safely stop background daemons, unregister OS startup items, and remove binary executables.
Deleting binary executable files alone will not kill background watcher daemons or unregister OS startup services for active projects. Before uninstalling, run dmai disable --all to cleanly terminate all background daemons, release SQLite locks, and unregister launch items. You can also run dmai projects to view all registered project paths and disable them individually.
Automatically runs `dmai disable --all`, unlinks `dmai` CLI, and removes system files.
curl -fsSL https://raw.githubusercontent.com/DevMemory-AI/devmemoryai/main/uninstall.sh | bashPaste in PowerShell to run `dmai disable --all`, unlink `dmai`, and remove global directories.
iwr -useb https://raw.githubusercontent.com/DevMemory-AI/devmemoryai/main/uninstall.ps1 | iex