AI Models
Rypl runs on the latest OpenAI models and upgrades them automatically, on every plan. Plans are not tied to model names: a model name on a plan card would go out of date the next time the model improves, and you would have to change plans to get the better one.
What differs by plan
| Simple | Pro | Ultra | |
|---|---|---|---|
| AI for replies | Smart AI | Smart AI | Most powerful AI, for the toughest questions |
| Reads the post (image, video, copy) before replying | — | ✅ | ✅ |
| Conversation memory | — | — | ✅ |
Everything else that uses AI is the same on every plan: intent detection, spam and hate detection, replies from your knowledge base, and replies in the commenter’s language.
What the AI does
- Reads what each comment wants. Intent detection labels questions about price, stock or shipping, complaints, praise and more. See Supported Intents.
- Spots spam and hate. Spam is flagged in about 1.7 seconds on average and hidden within seconds (RYPL’s own measurement).
- Writes replies in your brand voice from your knowledge base, in the language the person wrote in.
- Looks at the post or ad (image, video, copy) before replying, from Pro.
- Follows rules you write in plain words, from Pro: moderation rules and reply rules.
- Remembers earlier conversations with the same person, on Ultra.
Answers are only as good as the facts in your knowledge base. Keep prices, stock and policies current, and read replies closely in your first week.
Next Steps
- Pricing: what each plan includes
- Knowledge base: give the AI the facts it answers from
- Sign up: start a 7-day free trial
More on rypl.is
- Why generic auto-replies hurt your brand: what context-aware replies change
- AI vs rule-based comment automation: what actually works on Facebook
| Back to Reference | Back to Documentation | Visit Rypl |