Moderate a live chat

Use this when your chat needs a complete safety loop: users can report content, moderators can review it, decisions change room state, and connected clients reflect those changes.

What you implement: a report control in the chat client and a separate moderator surface on your trusted backend.

Safety workflow

sequenceDiagram
    actor U as User
    participant Chat as Chat client
    participant API as TalkLabs API
    participant Mod as Moderator console

    U->>Chat: Report message or user
    Chat->>API: Submit report
    API-->>Chat: Report accepted
    Mod->>API: List moderation queue
    API-->>Mod: Flagged messages and users
    Mod->>API: Approve or reject
    API-->>Mod: Decision applied
    Mod->>API: Optional mute, shadow ban, bounce, or purge
    API-->>Chat: Replace, remove, purge, or user-effect event
    Chat-->>U: Update visible room state

Map the two surfaces

User-facing chat

Control	Operation
Report message	Report a message
Report user	Report a user in this room
Refresh current effects	List room user effects

Confirm that a report was received. Do not tell the reporter that content was removed until a moderation decision or event update says so.

Moderator console

Moderator task	Operation
Review flagged messages	List messages needing moderation
Approve or reject a message	Apply message decision
Review flagged users	List flagged users
Apply a user decision	Apply user decision
Temporarily remove a user	Bounce user
Prevent visible posting	Shadow ban user
Temporarily stop posting	Mute user
Remove a user’s room messages	Purge user messages

Build the queue around decisions

A useful queue row contains the reported content, the author, report reason, room context, report count, and available decisions. Keep the decision action small:

export async function decideMessage(eventId, approve) {
  const response = await fetch(
    `https://api.sportstalk247.com/api/v3/${APP_ID}` +
      `/chat/moderation/queues/events/${eventId}/applydecision`,
    {
      method: 'POST',
      headers: {
        'x-api-token': API_TOKEN,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify({ approve }),
    },
  );

  if (!response.ok) throw new Error(`decision failed: ${response.status}`);
  return response.json();
}

After a decision succeeds, remove the item from the active queue or refresh that page. Do not infer success from a button click.

Reflect decisions in the room

Clients should already be processing the room updates stream. Route moderation events through the same reducer:

replace updates edited or moderated message content.
remove hides or tombstones a message.
purge removes the affected user’s messages from local state.
mute, bounce, and related events update the affected user’s controls and connection state.

If a client reconnects after missing these events, refresh room history or the affected state instead of assuming its cached view is current.

Moderator UX checklist

Show the TalkLabs room ID and your custom room ID as separate labeled values.
Make approve and reject visually distinct and confirm destructive bulk work.
Record the acting moderator in your own audit context where required.
Disable a queue action while it is pending.
Treat an already-decided item as a refresh condition, not an unknown failure.
Keep report reasons and user effects visible without exposing private tokens.