diff --git a/justfile b/justfile
new file mode 100644
index 0000000..57eb14f
--- /dev/null
+++ b/justfile
@@ -0,0 +1,78 @@
+# Vessel development commands
+# Run `just --list` to see all available commands
+
+# Default backend port for local development (matches vite proxy default)
+backend_port := "9090"
+
+# Default llama.cpp server port
+llama_port := "8081"
+
+# Models directory
+models_dir := env_var_or_default("VESSEL_MODELS_DIR", "~/.vessel/models")
+
+# Run backend locally on port 9090 (matches vite proxy default)
+backend:
+    cd backend && go run ./cmd/server -port {{backend_port}}
+
+# Run frontend dev server
+frontend:
+    cd frontend && npm run dev
+
+# Start frontend + backend in Docker
+dev:
+    docker compose -f docker-compose.dev.yml up
+
+# Start Docker dev environment in background
+dev-detach:
+    docker compose -f docker-compose.dev.yml up -d
+
+# Stop Docker dev environment
+dev-stop:
+    docker compose -f docker-compose.dev.yml down
+
+# View Docker dev logs
+dev-logs:
+    docker compose -f docker-compose.dev.yml logs -f
+
+# List local GGUF models
+models:
+    @ls -lh {{models_dir}}/*.gguf 2>/dev/null || echo "No models found in {{models_dir}}"
+
+# Start llama.cpp server with a model
+llama-server model:
+    llama-server -m {{models_dir}}/{{model}} --port {{llama_port}} -c 8192 -ngl 99
+
+# Start llama.cpp server with custom settings
+llama-server-custom model port ctx gpu:
+    llama-server -m {{models_dir}}/{{model}} --port {{port}} -c {{ctx}} -ngl {{gpu}}
+
+# Start Docker dev + llama.cpp server
+all model: dev-detach
+    just llama-server {{model}}
+
+# Check health of all services
+health:
+    @echo "Frontend (7842):"
+    @curl -sf http://localhost:7842/health 2>/dev/null && echo "  OK" || echo "  Not running"
+    @echo "Backend (9090):"
+    @curl -sf http://localhost:9090/health 2>/dev/null && echo "  OK" || echo "  Not running"
+    @echo "Ollama (11434):"
+    @curl -sf http://localhost:11434/api/tags 2>/dev/null && echo "  OK" || echo "  Not running"
+    @echo "llama.cpp (8081):"
+    @curl -sf http://localhost:8081/health 2>/dev/null && echo "  OK" || echo "  Not running"
+
+# Run backend tests
+test-backend:
+    cd backend && go test ./...
+
+# Run frontend type check
+check-frontend:
+    cd frontend && npm run check
+
+# Build frontend for production
+build-frontend:
+    cd frontend && npm run build
+
+# Build backend for production
+build-backend:
+    cd backend && go build -o vessel ./cmd/server