Disqus

Read and manage forums, threads, posts (comments), and users on the Disqus commenting platform. List and moderate comments, create threads and posts, and look up user activity via the Disqus API 3.0.

18 actions available

A scheduled sweep of the comment moderation queue begins

Agent pulls pending comments with List Posts filtered by moderation state

Agent scores each comment for spam and policy signals

Agent marks clear spam with Mark Post As Spam

Agent separates borderline comments that rules cannot settle

Agent posts each borderline comment to the moderation channel with context

A moderator decides Approve Post or Remove Post on the judgment calls

What This Integration Enables

Agents can run a comment pipeline that separates the mechanical from the judgment-bound. List Posts and List Forum Posts pull the queue, Mark Post As Spam and Remove Post handle the obvious cases, and Approve Post, Highlight Post, Close Thread, and Open Thread give the agent the full moderation surface to act on once a human decides. Create Thread opens a discussion for each new page, and Get User Details plus List User Posts let an agent pull a commenter's history when a call depends on context. These are automation exceptions: the items in an automated process that need human judgment when the rules do not cleanly apply, the data is incomplete, or the cost of being wrong is too high to automate. The connector hands those to a person and clears the rest.

Without FlowRunner

Scroll the whole queue Every comment gets read by a person, most of them spam or trivially fine

Slow visibility Legitimate comments wait behind the noise for a moderator to reach them

Inconsistent calls Removal decisions depend on which moderator happened to see the comment

With FlowRunner

Exceptions only The agent clears clear spam so the moderator sees only what needs a human

Faster surfacing Borderline comments arrive in chat with context the moment they land

One decision path Every judgment call routes through the same approve-or-remove step

Use Case Scenarios

A moderation queue that surfaces only the hard cases

A scheduled sweep calls List Posts filtered to the unapproved state. The agent scores each pending comment, marks unambiguous spam with Mark Post As Spam, and sets the clearly-fine comments aside for approval. The comments that rules cannot settle, an angry-but-legitimate complaint, a borderline promotional link, go to the moderation channel with the comment text and the author's history attached. A moderator picks Approve Post or Remove Post, and the agent applies it. The queue stops being a scroll and becomes a short list of real decisions.

A thread for every published page

When a new article is published, the agent calls Create Thread against the forum short name so the page opens with its discussion ready. If a discussion needs to be paused, Close Thread stops new comments while keeping the existing ones visible, and Open Thread reopens it later. The site stops shipping pages that have no comment thread wired up.

Featuring the best of a discussion

An agent watches an active thread with List Thread Posts and flags comments that read as high-signal. It drafts a shortlist to Highlight Post as featured comments and routes the shortlist to the community owner. Nothing is promoted without a person's nod, so the featured comments reflect an editorial call, not a heuristic.

Human-in-Loop Highlight

Removing a comment and marking a real person's post as spam are reputational acts, so the connector gates them behind a person. This is human-in-the-loop: an execution pattern where AI agents pause autonomously, assemble the relevant context and the decision choices available, route to a human via their preferred channel, and resume the moment the human responds. For every borderline comment, the agent posts to the moderation channel: "Needs a call on [page]: [comment text]. Author history: [summary]. Approve Post, Remove Post, or Mark Post As Spam?" The moderator decides, and the agent applies exactly that action through the Disqus API. Clear spam is cleared without a human; contested comments never are. That split is the whole value. Deleting the ninety obvious spam comments is not the work. The ten that need a human to weigh tone, context, and the cost of a wrong removal are the work, and this is the layer that owns them: a system above the comment platform that listens to the queue, gathers the author context the moderator needs, and pulls a person in only when being wrong would cost something. Deciding which items deserve that pause, across every channel where a brand faces the public, is an orchestration problem, and orchestration as a service is the category built for it. FlowRunner is built for that layer.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via preferred channel

Human decides

Agent resumes with decision

Agent Capabilities

18 actions

Forums

Get Forum Details Retrieves details of a Disqus forum (site) by its short name, including settings and subscriber and post counts. Used to confirm the target forum before acting on it.
List Forum Categories Lists the optional categories defined within a forum. Used when placing a new thread into a category.
List Forum Threads Lists threads (discussions) belonging to a forum, most recent first, with pagination. Used to audit or report on a forum's activity.
List Forum Posts Lists posts (comments) across all threads in a forum, most recent first, with moderation-state filtering. Used to export activity or feed a moderation dashboard.

Threads

Get Thread Details Retrieves a thread by numeric thread ID or by forum plus link (the page URL). Used to resolve the thread attached to a specific page.
List Thread Posts Lists the comments belonging to a single thread. Used to read one page's discussion for review or highlighting.
Create Thread Creates a new thread within a forum, typically one per page on your site. Requires a valid access token.
Close Thread Closes a thread so no new comments can be posted while existing ones stay visible. Requires a moderation access token.
Open Thread Reopens a previously closed thread so comments can be posted again. Requires a moderation access token.

Posts

Get Post Retrieves a single comment by its ID, including message, author, thread, and moderation flags. Used to pull the full context for a review decision.
List Posts Lists comments in a forum filtered by moderation state, most recent first. The primary action for pulling a moderation queue.
Create Post Creates a new comment on a thread, optionally as a reply. Requires a valid access token.
Approve Post Approves a comment, making it publicly visible and removing it from the queue. Applied after a moderator's decision. Requires a moderation access token.
Remove Post Removes a comment so it is no longer displayed. A moderation delete, applied only on a human's call. Requires a moderation access token.
Mark Post As Spam Flags a comment as spam, hiding it and training the Disqus spam filter. Used by the agent for clear spam and by a moderator for contested cases. Requires a moderation access token.
Highlight Post Highlights a comment as a featured comment in the thread. Used after a human approves the shortlist. Requires a moderation access token.

Users

Get User Details Retrieves the public profile of a user by numeric user ID, including reputation and aggregate counts. Used to give a moderator context on a commenter.
List User Posts Lists comments authored by a specific user across visible forums. Used to audit a user's activity before a moderation call.

Start building with Disqus

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us