API to Check if a DM Can Be Sent (Bot or User Account) – Avoid Unnecessary 403s

[lorenzo132]

Description

Currently, when a bot or normal user account tries to DM someone who has DMs blocked, the Discord API returns a 403 Forbidden (discord.Forbidden in discord.py). There is no way to check ahead of time whether a DM is allowed, forcing developers and users to rely on exception handling whenever a DM is sent.

Example Scenario (Bot):

try:
    await user.send("Hello!")
except discord.Forbidden:  # 403 Forbidden from API
    print("Cannot send DM to this user")

This works, but it’s inefficient and noisy for bots or scripts that need to check DM capability frequently.

Proposed Feature:
Add an API endpoint or property to check if a DM can be sent before attempting to send it, for both bot accounts and normal user accounts. Example usage:

if await user.can_dm():
    await user.send("Hello!")
else:
    print("Cannot DM this user")

Benefits:

• Reduce reliance on exception handling for normal flow control.
• Improve developer experience and clarity in bot or user scripts.
• Avoid unnecessary API calls that create DM channels only to fail.

Notes:

• Would prevent predictable 403 Forbidden errors in DM scenarios.
• Could be implemented as a new property on the User object or a dedicated API endpoint accessible to both bot and user accounts.

Steps to Reproduce

• Block a bot or user.
• DM the user that blocked you or your bot.
• See 403 appear.

Expected Behavior

• Block a bot or user.
• DM the user that blocked you or your bot.
• See a api response appear without getting a unwanted 403.

Current Behavior

• Block a bot or user.
• DM the user that blocked you or your bot.
• See 403 appear.

Screenshots/Videos

Client and System Information

• canary 471928 (bd7cff8) Host 1.0.768 x64 (72300) Build Override: N/A Windows 11 64-bit (10.0.26100)
• stable 471383 (a5dffa5) Build Override: N/A (Web - Windows)
• Android - 307.5 (4886) - canaryRelease

Why?

Even if the efficiency isn’t a huge win, it’s more about predictable behavior and avoiding unnecessary noise in production environments, especially for bots that handle many users.

[AlmostSuspense]

exposing this would let bad actors target only people with dms enabled, which would reduce the amount of requests they send and probably make it easier to go under the radar, i assume that's why they've never added this

[Soheab]

This could be part of the Create DM endpoint, which I just found always returns the channel with minimal information about the user.

This endpoint could return a 400 if the user's dm is closed and 403 if blocked (like sending currently). Maybe even specific JSON codes to determine whether the user doesn't share a server/app or not open in general.

[AlmostSuspense]

that wouldn't make sense because creating the dm channel isn't failing in those cases, it's sending to it that fails. you also need a successful response from that endpoint to get the dm channel id in case you need it to fetch messages sent in the past for example, and in that case it doesn't matter if the user has closed their dms

[advaith1]

Hi! This has previously been requested and denied in #612 - it's better to always send one request than to often send two requests.