Hi X team,
I’m building a web app that helps people manage conversations from their own X accounts, and I want to make sure our reply workflow follows the current rules before we launch.
**App ID:** 33168355 (Pay Per Use)
**API setup:** X API v2 with OAuth 2.0 user context. We use direct HTTPS requests from our backend and only the official X API—no scraping or browser automation.
The relevant endpoints are:
`POST /2/tweets` for public replies
`POST /2/dm_conversations/:dm_conversation_id/messages` for DM replies
X Activity API subscriptions for incoming mentions, replies, quotes, and DMs
Here’s how the feature works:
A customer connects their own X account. When they receive a mention, reply, quote, or DM, it appears in the web app. AI can suggest a draft response, but nothing is ever sent automatically. The customer sees the full draft, can edit or discard it, and must explicitly click **Send**. Only then does our backend make the X API request.
We also enforce these limits:
Replies are only available for incoming interactions involving the connected account.
Only one reply can be sent for each incoming interaction.
We do not scan timelines or keywords to contact people.
There are no unsolicited mentions, bulk replies, scheduled replies, or autonomous posting.
Public replies cannot mention people outside the existing conversation.
DMs are only answered within conversations initiated by the other person.
Likes and hide actions require their own explicit human click.
X content is not used to train an AI or ML model.
I found [this recent response from X staff](Approval request: AI-powered mention-triggered fantasy sports reply bot (@TheCallFF)), which says a compliant mention/reply/DM-triggered AI reply bot does not need additional written approval when those safeguards are enforced in software.
Since our workflow is human-reviewed and always requires the customer to click Send, can you please confirm that it also does not require separate approval?
If approval is still required, where should I request it now that the Policy Support form is no longer available?
Thanks!
2 posts - 2 participants
Read full topic
| # | Наименование новости | Тональность | Информативность | Дата публикации |
|---|---|---|---|---|
| 1 | Approval request: AI-powered mention-triggered reply bot (@MarlowStudent, App ID 32973549) | 0 | 9.98 | 30-08-2026 |
| 2 | Approval route requested: owner-reviewed AI-assisted X reply preparation | 0 | 10.42 | 27-08-2026 |
| 3 | X APIを利用した有料LINE・Discord案件検索通知サービスの利用可否について | 0 | 16.06 | 28-08-2026 |
| 4 | Developer Policy clarification for public-Post research | 0 | 6.14 | 27-08-2026 |
| 5 | API authorization issue OAuth2 | 0 | 8.78 | 28-08-2026 |
| 6 | Credits added to wrong developer account — which app/account should I use? | 0 | 6.39 | 27-08-2026 |
| 7 | OAuth 2.0 regression / authorization rejects numeric-only state values | 0 | 9.79 | 27-08-2026 |
| 8 | X Can't post link | 0 | 9.12 | 30-08-2026 |
| 9 | OAuth 2.0 PKCE authorization fails after user approval before callback | 0 | 7.24 | 28-08-2026 |
| 10 | Musk’s X poses “serious risk to Americans’ privacy,” advocates warn FTC | -2 | 7 | 02-07-2026 |