Self-Hosted
Install and run Memtrace on your own infrastructure.
Prerequisites
- Go 1.21+
- A running Arc instance (port 8000)
Build from Source
git clone https://github.com/basekick-labs/arc-memory.git
cd arc-memory
make build
The binary will be in ./memtrace.
Configuration
Create a configuration file:
cp memtrace.toml memtrace.local.toml
Edit memtrace.local.toml with your Arc URL and credentials. See the Configuration reference for all available options.
Minimal configuration:
[arc]
url = "http://localhost:8000"
api_key = "your_arc_api_key"
database = "memory"
measurement = "events"
[auth]
enabled = true
db_path = "./data/memtrace.db"
Run
./memtrace
Or specify a custom config file:
./memtrace -config /path/to/memtrace.toml
On first run, Memtrace prints your admin API key. Save it — it's shown only once.
FIRST RUN: Save your admin API key (shown only once)
API Key: mtk_...
Systemd Service
Create /etc/systemd/system/memtrace.service:
[Unit]
Description=Memtrace Memory Layer for AI Agents
After=network.target
[Service]
Type=simple
User=memtrace
Group=memtrace
WorkingDirectory=/opt/memtrace
ExecStart=/usr/local/bin/memtrace -config /etc/memtrace/memtrace.toml
Restart=on-failure
RestartSec=5s
[Install]
WantedBy=multi-user.target
Enable and start:
sudo systemctl enable memtrace
sudo systemctl start memtrace
sudo systemctl status memtrace
System Requirements
Minimum
- 1 CPU core
- 512 MB RAM
- 1 GB disk space
Recommended
- 2+ CPU cores
- 2 GB RAM
- 10 GB disk space (for SQLite metadata/auth database)
Memtrace is stateless for memory data (stored in Arc), but maintains SQLite databases for auth, agents, and sessions.
Ports
- 9100: HTTP API (default)
Configure via server.port in the config file or MEMTRACE_SERVER_PORT environment variable.
Verify Installation
Check health:
curl http://localhost:9100/health
Response:
{"status": "ok", "service": "memtrace", "uptime": "2h30m15s"}
Check Arc connectivity:
curl http://localhost:9100/ready
Response:
{"status": "ready", "arc": "connected"}
Next Steps
- Read the Configuration reference
- Explore the API Reference
- Try the Quick Start guide