AI replies and moderation for Facebook, Instagram and TikTok ad comments
AI support agent for B2B software companies
Stage
Startup
MVP
Founded
2025
Not given
Based in
IS
Not given
Pricing model
Free trial
Contact for pricing
Pricing
Simple $19/month, Pro $49/month, Ultra $89/month. Yearly billing about 20% off: Simple $15/month ($182/year), Pro $39/month ($470/year), Ultra $71/month ($854/year). 7-day free trial on every plan.
USD 299 per month with 1,000 conversations included, then USD 0.30 for each additional conversation. No seat or API fees. Free preview for setup and testing. 30-day first-month refund guarantee.
Key features
AI replies to comments and DMs on Facebook and Instagram
AI replies and moderation for TikTok ad comments
Knowledge base from website, files and Q&A on every plan
Spam, hate and banned-word filters on every plan
Intent detection on every comment
Moderation rules in plain words with Pro and Ultra
Replies in commenter's language
No limit on comments processed on any plan
Conversation memory on Ultra
Reads the post or ad before replying with Pro and Ultra
Completes approved support tasks like billing and plan changes
Verifies customer identity before accessing accounts
Shows exact change and cost before customer confirms
Confirms success through connected systems before reporting completion
Automatically detects and re-tests answers when documentation changes
Hands off uncertain cases to human support with context
Early-access fit review confirms sources and task connections before go-live
Website widget and API integration options
Daily morning note of handled requests and team review
What makes it different
38% of ad comments come from people about to buy
Replies on-brand within 60 seconds
Spam spotted in 1.7 seconds on average and hidden within seconds
No password required, uses official Meta and TikTok APIs
About 10% ROAS lift in first month
Requires source documentation and approved workflows before activation, never infers policies
Automatically monitors source changes and re-tests affected answers
Data never used to train models, with strict safeguards on material claims
Accepts up to five new teams per month with personalized review
RYPL
Sondel