Skip to content

Latest commit

 

History

History
78 lines (54 loc) · 2.16 KB

File metadata and controls

78 lines (54 loc) · 2.16 KB

Runtime installation

Install the runtime daemon on every machine that should execute Odoo agent work.

Quick path

  1. Create a runtime record in Odoo.
  2. Generate and copy its API key.
  3. Clone this repository on the execution machine.
  4. Run the installer.
  5. Choose manual mode first.
  6. Confirm heartbeat in Odoo.
  7. Switch to service/daemon mode only after the manual run works.

Linux

git clone https://github.com/nicolasramos/odoo-agent-runtime.git
cd odoo-agent-runtime
bash install.sh

Manual run:

python3 daemon.py

Systemd mode is available from the installer when selected.

macOS

git clone https://github.com/nicolasramos/odoo-agent-runtime.git
cd odoo-agent-runtime
bash install.sh

Launchd mode is available from the installer when selected.

Windows

git clone https://github.com/nicolasramos/odoo-agent-runtime.git
cd odoo-agent-runtime
.\install.ps1

Scheduled Task mode is available from the installer when selected.

Required values

Value Example
Odoo URL https://odoo.example.com
Runtime API key odoo_rt_xxx
Runtime name agent-worker-01
Poll interval 10

Optional Odoo database

ODOO_DATABASE is optional for a single-database Odoo deployment. Provide it when the Odoo URL serves multiple databases, or when your runtime must always target one specific database. On Linux and macOS, enter it at the install.sh prompt; on Windows, enter it at the install.ps1 prompt. Each installer writes an empty value when it is not needed.

Keep the runtime API key separate from the Odoo URL. Enter it only at the Runtime API key prompt; never append it to a URL, query string, bookmark, or command history.

Existing .env

The installer does not silently overwrite .env. If .env exists, it asks before replacing it. If you decline, it writes .env.new and does not install/start a background service with stale settings.

First validation

After starting the daemon:

  1. Open Odoo.
  2. Go to AI Agents → Runtimes.
  3. Confirm Last Seen updates.
  4. Send a simple task to an agent assigned to that runtime.
  5. Confirm logs and final status.