Install memory
where your agents live.
One guided installer detects your machine, offers local or cloud setup, verifies AutoMem, then reviews every Codex, Claude Code, Cursor, OpenClaw, and Hermes write before installing.
curl -fsSL https://automem.ai/install.sh | shHosted Cloud
Deploy on InstaPods or Railway, paste the generated HTTPS endpoint and token, then let the wizard wire your local agents.
curl -fsSL https://automem.ai/install.sh | AUTOMEM_INSTALL_TARGET=cloud shLocal Docker
Clone AutoMem into ~/.automem/server, generate local tokens, start Docker Compose, verify /health, then install agent rules.
curl -fsSL https://automem.ai/install.sh | AUTOMEM_INSTALL_TARGET=local shExisting Endpoint
Use a running AutoMem instance you already control. The wizard verifies it before touching agent configuration.
curl -fsSL https://automem.ai/install.sh | AUTOMEM_INSTALL_TARGET=existing AUTOMEM_API_URL=https://memory.example shReview before write.
The installer writes environment values, then delegates to the existing platform installers. Config files are backed up before changes, and platforms with unreliable automatic config still print exact manual snippets.
| Flag / Env | Purpose |
|---|---|
| AUTOMEM_INSTALL_TARGET | local, cloud, or existing |
| AUTOMEM_CLIENTS | Comma-separated agent list |
| AUTOMEM_API_URL | AutoMem HTTP API endpoint |
| AUTOMEM_API_KEY | Bearer token for authenticated deployments |
| AUTOMEM_LOCAL_DIR | Local backend checkout path |
| AUTOMEM_YES=1 | Apply the reviewed plan without prompting |
| AUTOMEM_DRY_RUN=1 | Print the plan without writing files |
| AUTOMEM_NO_AGENT_INSTALL=1 | Configure endpoint only |