Every comment answered. Even at 3am.
A question left sitting for nine hours is a sale gone elsewhere. Mediux reads every comment and DM, decides whether it should answer, and replies in your brand's voice within seconds.
Free plan · No credit card required · Cancel any time
Auto-Reply · Instagram
✦ AI reply · your brand voice · sent automatically
Works across every channel you publish to
0/7
Always on
Including while you sleep
0
Variations per rule
Rotated at random
0
Networks
Comments and DMs
0
Inbox
Every channel in one place
Scoped, written, then watched
Switch between the three parts of a rule below.
Decide exactly what gets answered
Target a specific post, a keyword, or an intent like pricing questions. Everything else stays untouched in the inbox for you. Automation you cannot scope is automation you cannot trust.
Active rules
Inquiry
DMs · Reply to all
Data Viz
Comments · keywords
Price · cost
Comments
Automation that does not sound automated
Three things stop this reading like a bot — and one stops it going wrong.
Narrow by post, keyword or intent
You choose what qualifies. A rule can watch one campaign post, fire only on the word 'price', or route anything that reads as a demo request. Everything outside the rule is left alone.
- Rules per post, per keyword, or per detected intent
- Separate handling for comments and DMs
- A blacklist for words that should never trigger a reply
Active rules
Inquiry
DMs · Reply to all
Data Viz
Comments · keywords
Price · cost
Comments
One place for every conversation
Comments and DMs from every connected channel land in a single feed, with what was auto-answered clearly marked and what needs a human waiting at the top. Take over any thread by hand at any point.
- Instagram, Facebook, X and Threads in one stream
- Auto-replied, liked, hidden and pending clearly separated
- Jump in and reply by hand whenever it matters
Recent activity
@sarok
Auto-replied to a comment
just now
@farhan
DM answered · Greeting
2m
@meta_user
Comment hidden · spam
5m
Built so nothing goes out that you would not have sent
Approval mode holds replies for a human. A blacklist catches the words that should always escalate. And a runtime content check blocks anything that would break a platform's reply policy before it posts.
- Approval mode for the first week of any new rule
- Blacklist words that must always reach a person
- Policy check on every reply before it is sent
Unified inbox
Everything you can automate
And everything you can stop it touching.
Comment replies
Answer under the post, where everybody else considering the same question can read it too.
DM replies
Private answers to private questions.
Keyword triggers
Fire only when a message means what you meant.
Intent detection
Pricing, shipping, a demo request — each handled apart.
Blacklist
Words that must never trigger an automatic reply.
Spam hiding
Hide the junk instead of replying to it.
Approval modeSafe start
Queue replies for a human to release.
Review replies
Google Business reviews answered in your voice, within a star range you choose.
Up and running in three steps
Step 01
Pick what to watch
A post, a keyword or an intent. Start narrow — one campaign is a good first rule.
Step 02
Generate the replies
The AI writes a set of variations in your brand voice. Edit any of them before they go live.
Step 03
Approve, then let go
Run it in approval mode for a week. When you trust it, switch it to automatic and stop watching.
The questions do not wait for office hours
Most of them arrive when nobody is looking — and most of them are the same five questions.
The difference it actually makes
Answering by hand
- A question posted at 11pm answered at 9am, if at all
- The same five questions retyped every single week
- Comments buried under a week of notifications
- Nobody sure whether that DM was already handled
With auto-replies on
- Answered in seconds, whatever time it arrives
- Twenty rotating variations, so it never reads as canned
- One inbox across Instagram, Facebook, X and Threads
- Anything unusual held back and flagged for a person
The fastest reply usually wins the sale — and it is rarely the best-written one.
Anything the rules do not confidently cover waits for a human. That restraint is the feature.
Questions, answered
Will people be able to tell it is automated?
The giveaway is always the same canned sentence appearing under every comment — which is exactly why rules generate up to twenty variations and rotate them. Replies are written in your saved brand voice, and you can edit every variation before the rule goes live.
How many auto-replies do I get?
Pro and Agency are unlimited, and sending costs nothing extra — the AI writes the variations once and the rule reuses them. The Free plan includes 4 sends across comments and DMs as a lifetime trial.
What if someone asks something the rules do not cover?
Nothing happens automatically. It stays in the inbox marked as needing a human, which is the right outcome — an automation that guesses at an unfamiliar question does more damage than one that waits.
Can I stop it replying to angry comments?
Yes. Keep rules scoped to the intents you want handled, and use the blacklist for words that should always route to a person. For Google Business reviews you can also set a star range, so only the reviews you choose ever get an automatic response.
Which platforms does it cover?
Comments and DMs on Instagram, Facebook, X and Threads, plus Google Business Profile review replies. Each runs through the platform's official API with tokens you can revoke at any time.
Can I approve replies before they send?
Yes. Approval mode queues every generated reply for a human to release, which is the sensible way to run the first week of any new rule before you let it go on its own.
Stop losing leads to a slow reply.
Set one rule on one post and watch what happens. Free to start, no card required.