Live — connect today
FounderReply drafts X posts and replies and holds them for approval before anything sends from your connected account. X is Pro-only and bring-your-own-credentials: you register your own X developer app, so the API tier, the rate limits and the spend stay yours rather than being pooled across customers.
On X
Replies and posts, on your own X developer app. Pro plan, bring your own keys.
Recent-search finds the posts worth answering, and each draft carries the parent post so you approve a reply that reads as part of the conversation.
Recent search needs at least X’s Basic API tier; on the Free tier the search call is unentitled and discovery simply returns nothing rather than failing loudly.
Publishing your own content and replying on your own posts can be dialled up to automatic. Messaging strangers and commenting on their posts cannot.
X’s automation rules permit automated posting of your own content and treat bulk automated replies and unsolicited DMs as spam — which is exactly where the cap table draws its line.
X is the one platform where you supply the developer credentials. That keeps your account under your own API agreement instead of a shared app that everyone’s behaviour affects.
X is the only entry in the bring-your-own-credential list, and the connect route rejects the attempt unless the workspace is on Pro and has its own client credentials stored.
The room
X is where product launches, build-in-public threads and "what are you using for X?" questions happen in the open, and where a good reply is more visible than a post. The people you want are already in someone else’s replies. That makes X the platform where drafting speed matters most and where a badly judged automated reply is punished fastest.
X’s spam policy is written around patterns rather than intent: identical replies at volume, unsolicited DMs, and reply-guy behaviour on accounts you have no relationship with. Everything that keeps you inside the rules also happens to be what makes a reply land.
The ceiling
Two layers decide whether an action happens by itself: your autonomy dial, and a hard ceiling we impose per platform. The safer of the two always wins, so the dial can make things stricter and never looser. This table is generated from that ceiling.
| Action | Ships as | Most permissive allowed |
|---|---|---|
| Publish your own postOriginal content from your connected account | Waits for your approval | May run automatically |
| Reply on your own postsAnswering comments left on your content | Waits for your approval | May run automatically |
| Direct messagePrivate messages to a person | Waits for your approval | Waits for your approval |
| Comment on someone else’s postThe highest-risk action on every platform | Not permitted | Waits for your approval |
X’s Automation Rules permit automated posting of your own content but prohibit bulk automated replies and unsolicited automated DMs. Own-content posting and replies on your own posts may therefore run unattended; DMs and comments on other people’s posts are held for a human.
Authorisation happens against your own X app with PKCE, and the offline.access scope is what supplies a refresh token so the connection survives without you re-authorising.
Nothing here is a marketing paraphrase — it is the string the connect flow sends, copied from lib/connectors/x.ts.
tweet.read tweet.write users.read dm.read dm.write like.write offline.accessFAQ
Not per se. X permits automated posting of your own content and prohibits bulk automated replies and unsolicited automated DMs. FounderReply encodes that split directly: posting and replying on your own account may run automatically, while DMs and comments on other people’s posts always wait for a human.
Because X charges for API access by app, not by user, and its automation rules bind the app owner. Running every customer through one shared app would pool both the cost and the risk. Your own app keeps your rate limits, your tier and your compliance record yours.
Pro. The connect endpoint checks the tier before it checks anything else and refuses X on Free and Starter, then refuses again if the workspace has not stored its own X client credentials. Every other platform is available from the Free plan.
It can draft them, and it will never send one on its own. The DM ceiling on X is human approval and there is no setting that lifts it, because unsolicited automated DMs are the single behaviour X’s spam policy names most explicitly.
The free plan connects one account and drafts replies for you to approve. Nothing is published until you release it.