תוספות
על אודות תוספת זו
Outbound Guardian adds an AI-powered safety net to your email workflow. Before any email leaves your outbox, it runs your custom checks through an LLM and blocks the send if any issues are detected.
How it works
Key features
Requirements
Getting started
Privacy
Your email content is sent to the OpenAI API for evaluation using your own API key. No data is sent to the extension developer or any other third party. See the full privacy policy for details.
How it works
- You define check rules in plain language (e.g. "contains em-dash characters" or "tone is too informal for a business email")
- When you click Send, each rule is evaluated by the LLM against your email
- If any rule triggers, the send is cancelled and a warning dialog appears
- From the dialog you can:
- Edit — return to the compose window to fix things manually
- Auto-fix — let the LLM apply corrections based on your improvement instructions
- Send anyway — override the warning and send as-is
Key features
- Fully customizable rules — you write the conditions in plain language, the LLM evaluates them
- Optional auto-fix — add an improvement prompt to any rule and the LLM will rewrite the email for you
- Non-blocking on errors — if the API is unreachable or misconfigured, your email sends normally
- No data collection — your emails are sent only to your own API provider using your own key
Requirements
- Thunderbird 128 or later
- An OpenAI API key (requires a paid OpenAI account)
Getting started
- Install the extension
- Open Add-ons → Outbound Guardian → Options
- Enter your OpenAI API key and click Save Settings
- Click Add Rule to create your first check
- Compose an email and hit Send — Outbound Guardian takes it from there
Privacy
Your email content is sent to the OpenAI API for evaluation using your own API key. No data is sent to the extension developer or any other third party. See the full privacy policy for details.
