• Checkstep Docs Setup checklist Documentation - Setup Setup checklist Set up your moderation stack from policies to live content. Follow the steps below - most teams are up and running in under an hour. Estimated time: 45–60 minutes Before you begin Have these ready 2 min review You don't need everything on day one - but having these close will speed things up considerably. Your community guidelines or platform policy document - a rough draft is fine
  • A clear sense of your primary content types: text, images, video, audio, or a combination
  • At least one internal stakeholder who owns policy decisions
  • Your platform's API credentials - needed for integration, but you can come back to this later

Define your policies

Your Policy Dashboard is where you define what your platform stands for. Think of policies as your rules of the road - everything else in Checkstep flows from here.

  • Navigate to the Policy Dashboard
  • Create your first policy - e.g. Hate Speech, Spam, or Sexual Content
  • Add the public-facing policy text - this is what users see if content is removed
  • Add internal guidelines - definitions and edge cases to help moderators stay consistent
  • Save and publish your policy

Set up your AI models

Checkstep's AI Hub lets you connect the models that scan your content. Most teams start with 3–5 models. You only pay for scans on content eligible for each model - Checkstep can help forecast usage before you commit.

  • Go to Settings → Strategies
  • Browse available models and enable the ones relevant to your content types
  • Set confidence thresholds for each model using the guide below

Connect your policies to your AI

Policy Rules are where your policies and AI models meet. This is how Checkstep knows what action to take when a model returns a result on a piece of content.

  • Inside each policy, navigate to the Rules tab
  • Add a rule for each relevant AI label - e.g. "if hate speech confidence is above 95%, auto-enforce"
  • Set the action for each rule: auto-remove, send for review, or trust
  • Repeat for each policy you've created

Configure your Moderation Queue

Your Moderation Dashboard is where flagged content lands for human review. Set this up before content starts flowing so your team knows exactly what to do when items arrive.

  • Define your moderator actions - e.g. Enforce, Dismiss, Escalate, Suspend User - triggerable by shortcut key or in bulk
  • Configure queue routing - decide which content goes to which moderators
  • Enable moderator wellbeing settings: blurred imagery and greyscale for sensitive content
  • Set up Slack or email notifications for new flagged items Optional
  • Enable ModBot for AI-assisted decisions on clear-cut cases Optional

Run a test ingest

Before going live, run a small batch of test content through the system. This confirms your policies, models, and rules are all connected correctly - and gives your team a chance to see the workflow in action.

  • Go to Settings → Test & Integrate
  • Download the CSV sample template
  • Add a few rows of test content covering your main content types
  • Upload and run the ingest
  • Open your Moderation Dashboard - flagged items should appear
  • Verify the correct policies and labels are triggering on the right content

Set up compliance

If your platform is subject to the EU Digital Services Act or the UK Online Safety Act, Checkstep automates your compliance reporting so you don't have to manage it manually.

  • Go to Regulation Settings
  • Paste in your regulatory API key provided by the relevant authority
  • Confirm that moderation actions will auto-generate compliance reports
  • Review your Transparency Portal settings - this is what users see when their content is removed

You're live Your moderation stack is set up. Here's what to keep an eye on during your first week of live data.

Week 1 - what to monitor in your Reporting Dashboard

After 1–2 weeks of live data, book a threshold tuning session with your account manager.