Engineering Memory for Modern Software Development.
DevMemory AI is a local-first, zero-cloud engineering memory system designed to preserve software understanding, architectural decisions, component dependencies, and code evolution across development lifecycles.
Traditional chat histories and commit logs record what changed, but fail to capture why decisions were made, how component boundaries evolved, or what technical debt was introduced.
DevMemory AI compiles filesystem changes into Canonical Engineering Memory—a single source of truth that powers automated build documentation, interactive knowledge graphs, repository context, and local AI reasoning.
- Local-First & Offline: 100% private. All SQLite databases (
.devmemory/index.db), AST indexes, and AI queries run locally on your machine. Zero cloud telemetry. - Single Runtime Orchestration: Every interaction (CLI, Watcher daemon, Dashboard) routes through a single
Runtimeinstance (src/runtime/index.ts) ensuring deterministic state handling. - Intent-Driven Engineering Compiler: Synthesizes high-level engineering intent, architectural impact, decisions, tradeoffs, and testing guidance without raw filename listing or generic AI filler.
- Unified Canonical Engineering Model: Both human documentation (
Full Context) and AI working prompts (Project Regeneration Prompt) derive 100% equivalent engineering facts from the same canonical model. - Interactive Knowledge Graph & Dashboard: Local HTTP dashboard (
http://localhost:31415) featuring Cytoscape.js Knowledge Graph visualization, Mermaid build diagrams, timeline sliders, and settings management. - Adaptive Cooldown Watcher: Background daemon with adaptive windowing (20s–60s) to batch rapid file edits cleanly while ignoring
.devmemoryand internal infrastructure. - SQLite WAL Concurrency: High-throughput SQLite FTS5 index running in WAL mode with busy timeout handling.
Repository Filesystem (Chokidar Watcher)
│
▼
Transaction Engine (Adaptive Cooldown)
│
▼
Engineering Compiler (Intent & AST Parsing)
│
▼
Canonical Engineering Memory (SQLite WAL Index)
┌────┴──────────────────────────┐
▼ ▼
Web Dashboard (Port 31415) CLI & AI Query Engine
curl -fsSL https://raw.githubusercontent.com/DevMemory-AI/devmemoryai/main/install.sh | bashiwr -useb https://raw.githubusercontent.com/DevMemory-AI/devmemoryai/main/install.ps1 | iexThe installer automatically detects and sets up Git, Bun runtime, Ollama LLM engine, and links the global dmai CLI.
Alternatively, clone the repository and install dependencies using Bun:
git clone https://github.com/DevMemory-AI/devmemoryai.git
cd devmemoryai
bun install
bun link
### 2. Link CLI Binary
```bash
bun link
cd /path/to/your/project
dmai initdmai watch
dmai dashboard| Command | Description |
|---|---|
dmai init |
Initialize DevMemory AI in the current repository |
dmai watch |
Start the background repository watcher daemon |
dmai stop |
Stop the background repository watcher daemon |
dmai disable |
Safely stop processes, unregister OS launch services, & purge .devmemory/ for clean deletion |
dmai continue |
Resume a stopped repository watcher daemon |
dmai restart |
Restart the repository watcher daemon |
dmai status |
Display project health, watcher status, and build count |
dmai ask "question" |
Query engineering memory locally using AI |
dmai search "query" |
Search FTS5 engineering index and fingerprints |
dmai dashboard |
Open local HTTP engineering dashboard |
dmai config |
View or update configuration settings |
dmai logs |
View runtime logs |
dmai doctor |
Run comprehensive system health diagnostics |
dmai version |
Display DevMemory AI version information |
- Architecture Guide
- Installation Guide
- CLI Reference
- Configuration Reference
- Troubleshooting & FAQ
- Release Notes
- Contributing
- Code of Conduct
- Security Policy
DevMemory AI source code is publicly viewable for inspection, educational evaluation, and local execution under the DevMemory AI Source-Available License. Source code modification, derivative works, and commercial redistribution are strictly prohibited.
- Creator: Umesh Adabala (Student Engineer • AI Builder)
- Portfolio: https://umeshadabala.github.io
- GitHub: https://github.com/umeshadabala
- Repository: https://github.com/DevMemory-AI/devmemoryai
- Discord Community: https://discord.gg/ycF48ADnx
- Website: https://devmemoryai.org