Spam Filtering
Postbox offers two spam protection strategies. Enable spam protection on any form and choose the level that fits your needs.
Strategies
Section titled “Strategies”Standard protection (free)
Section titled “Standard protection (free)”Heuristic detection that catches common spam patterns. Includes honeypot field checking and content moderation. No AI credits needed - runs on every plan at no cost.
Enable by setting spam_protection_enabled: true with spam_protection_strategy: "standard" (the default).
Intelligent analysis (AI-powered)
Section titled “Intelligent analysis (AI-powered)”LLM-powered analysis that understands context, detects sophisticated spam, and provides confidence scores with human-readable explanations. Uses AI credits.
Enable by setting spam_protection_enabled: true with spam_protection_strategy: "intelligent".
How it works
Section titled “How it works”When a submission arrives with spam protection enabled:
- Standard checks run first - Honeypot fields, heuristic patterns, content moderation
- If intelligent is enabled - AI analyzes the content and assigns a confidence score
- Flagging - Spam is flagged silently. The submitter always gets a
201response. - Pipeline stops - If spam is caught, translation and smart replies are skipped.
Submission spam fields
Section titled “Submission spam fields”Every submission includes spam metadata:
| Field | Description |
|---|---|
spam_flag | true if flagged as spam, false otherwise |
spam_confidence | AI confidence score (0.0 to 1.0), or null if not AI-analyzed |
spam_reason | Human-readable explanation, or null |
Viewing spam
Section titled “Viewing spam”In the dashboard, filter submissions to show inbox (non-spam), spam only, or all. The API supports the same filtering via the filter query parameter.
| Strategy | Cost |
|---|---|
| Standard (heuristic + honeypot) | Free, always |
| Intelligent (AI-powered) | $0.005 per submission analyzed |
On the free plan, the 50 one-time AI credits cover intelligent analysis. When credits run out, AI features stop but standard protection continues working.
On Pro, you get 500 credits/month. After they’re used, AI features continue via metered billing.
Disabling spam filtering
Section titled “Disabling spam filtering”Toggle per-form in the form settings or via the API (spam_protection_enabled: false). Submissions are still accepted and rate-limited - they just won’t be analyzed for spam.