Migration Guide
9 min read

Community Sift alternative: replace it before your contract ends

Community Sift moderated chat for Xbox, Supercell, EA and many other game communities. Microsoft bought it with Two Hat in 2021, and it is now winding it down: existing customers have been told their contracts will be honored until they expire and then not renewed.

If Community Sift filters your game chat or usernames, you need a replacement live before your renewal date. This guide maps what Community Sift actually did to the alternatives, shows what each one costs, and walks through moving your chat to Paxmod, the real-time chat moderation API you can start on today without a sales call.

Your deadline is your contract end date

Microsoft has not published a shutdown date, and its product page is still up. The wind-down is handled customer by customer, so ask your Microsoft account team when your contract ends. Then export your policy guides, rules and word lists now, while you still have full access.1

The short answer

Game chat and usernames

Paxmod. Built for live game chat, self-serve, published pricing, and live the same day.

Voice chat

GGWP or Modulate. Paxmod does not moderate audio.

Video and broad UGC

Hive or Lasso. Paxmod covers text and images, not video.

What Community Sift did, and what replaces it in Paxmod

Community Sift was more than a word filter, so check every replacement against everything it did for you, not just its toxicity score.2 Here is how each capability maps to Paxmod, including where it works differently and where Paxmod has no equivalent.

Community SiftPaxmod
Topic classification on a sliding risk scale
13 categories, each scored 0 to 1 with its own threshold per project. Anything the fixed categories miss goes into an Advanced mode policy written in plain English.
Unnatural Language Processing (l33t speak, Unicode tricks, misspellings)
Every message is normalized before scoring: leetspeak, Unicode lookalike characters and hidden zero-width characters. Test your own evasion samples during the trial.
Policy guides per feature (strict public chat, relaxed clan chat, strictest usernames)
One project per surface, each with its own key, categories, thresholds, allow list and policy.
Username moderation
Send usernames to the same text endpoint, from a stricter usernames project.
Grooming and predator detection
Grooming-risk and personal information detection ("where do you live", moving a child to another app). Studio adds conversation context, so risk that builds over several messages is caught.
User Reputation (Trusted, Default, Not-Trusted)
Send user_id with each message. User Analysis keeps each player's flag history and a risk summary. Your game applies the trust level; Paxmod does not switch states for you.
Image moderation
Studio plan: POST /api/v1/image with 13 image categories, including Sexual/Minors and Violence/Graphic.
22 native languages
Language is detected automatically, with no language parameter. Test your top languages on real traffic during the shadow run.
Review queue and Predictive Moderation for player reports
The console lists flagged messages and per-user analysis, and lets you mark true and false positives. Paxmod does not triage player reports.
Video moderation
Not offered. Paxmod moderates text and images.
Hosted on Microsoft Azure
Enterprise plan: data residency in Canada or the EU, an SLA and invoicing.
Direct replacement Works differently Not offered

Community Sift alternatives compared

AlternativeContent typesBuilt for game chatPublished pricingStart without a sales call
PaxmodText, usernames, imagesYes: Unity, Roblox, Minecraft, DiscordYes, from $29/month for 50,000 messagesYes: test key in one API call, 14-day free trial
GGWPChat, voice, images, usernames, player reports3YesNo, quoted after a demoNo
Lasso ModerationText, images, video, audioGeneral purposeYes, from $99/month for 10,000 items4Yes
HiveText, images, video, audioNoText at $0.50 per 1,000 requests, on an annual contract5Developer credits for testing
CheckstepText, images, videoPartly, compliance-ledNoNo
CleanSpeakText, usernamesYes, list-based filteringNoNo

What it costs at game-chat volume

Community Sift never published a price. Here is what the alternatives that do publish one charge for a month of text moderation.

Monthly messagesPaxmodLassoHive (text only)
100,000$49 (Indie: $29 for 50,000, then $0.40 per 1,000)$499 (Pro, up to 100,000 items)About $50, annual contract
500,000$149 (Studio, includes image moderation and conversation context)Enterprise quote (above the Pro limit)About $250, annual contract

Paxmod figures use Standard mode (1 credit per message). Advanced mode, where messages are judged against your own written policy, uses 2 credits per message. GGWP, Checkstep, CleanSpeak and Community Sift quote prices through sales.

Why game studios pick Paxmod to replace Community Sift

  • Live this week, not next quarter. No procurement cycle. Your engineer (or their AI coding agent) can get a working key in one API call and moderate real messages in minutes.
  • Built for live game chat. About 500ms per message, more than 10 million messages moderated a month, and guides for Unity, Roblox, Minecraft and Discord.
  • The same child-safety focus. Two Hat was founded to protect kids online. Paxmod flags personal information requests and grooming-risk patterns, and our grooming detection research is built with the Australian Institute for Machine Learning. See our COPPA-oriented controls.
  • Your rules in plain English. Community Sift's policy guides were sliders over fixed topics. Paxmod keeps the per-category sliders and adds Advanced mode: write the rule ("flag attempts to trade accounts for real money") and every message is judged against it.
  • Hard to dodge. Leetspeak, Unicode lookalikes and hidden zero-width characters are normalized before scoring, so "h4t3" scores like "hate".
  • Prices you can see. Indie $29/month, Studio $149/month, Enterprise from $1,500/month, with data residency in Canada or the EU. No minimum annual contract to get started.

Mapping your Community Sift integration to Paxmod

In Community SiftIn Paxmod
The message text you submitmessage (up to 5,000 characters)
Player or account iduser_id, which powers per-user history and User Analysis
Room, channel or chat idcontext_id, which powers conversation context on Studio
Policy guide for a featureA Paxmod project. The API key you send selects it
Topic risk levelscontent_moderation.categories[name].score compared with that category's threshold
Allow or block decisionresult: "flagged" or "not_flagged", with a reason
Custom word listsAllow lists per project (skip, remove or replace a term), plus an Advanced mode policy for terms you want blocked

curl

curl -X POST "https://www.paxmod.com/api/v1/text" \
  -H "Authorization: Bearer $PAXMOD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "add me on snap, whats ur address",
    "user_id": "player_123",
    "context_id": "lobby_7"
  }'

Node (server side)

const res = await fetch("https://www.paxmod.com/api/v1/text", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    Authorization: `Bearer ${process.env.PAXMOD_API_KEY}`,
  },
  body: JSON.stringify({ message: text, user_id: playerId, context_id: roomId }),
});
const data = await res.json();
if (data.status === "error") throw new Error(data.error.code);

if (data.result === "flagged") {
  // Same place your Community Sift "reject" branch lived.
  hideMessage(data.reason);
}

Response

{
  "status": "success",
  "id": "5f0c2a8e-1d9b-4c47-9a51-3c1f6f2b7e10",
  "result": "flagged",
  "reason": "Content flagged for Personally Identifiable Information",
  "content_moderation": {
    "flagged": true,
    "categories": {
      "Personally Identifiable Information": { "flagged": true, "threshold": 0.5, "score": 0.91 },
      "Harassment": { "flagged": false, "threshold": 0.5, "score": 0.03 },
      "Sexual": { "flagged": false, "threshold": 0.5, "score": 0.02 }
    }
  }
}

Call Paxmod from your game server, never from the client, so your API key stays private. The API docs cover image moderation, errors and rate limits.

How to migrate from Community Sift in 6 steps

  1. 1

    Export your Community Sift setup

    While you still have access, export your policy guides, custom rules, word lists, per-language settings and a sample of messages Community Sift flagged and allowed. This is your requirements list and your test set.

  2. 2

    Get a Paxmod key

    Get a test key in one API call with no signup (POST https://www.paxmod.com/api/v1/agent-key, 100 messages), or start the 14-day free trial (10,000 messages, every feature) and create a project in the console.

  3. 3

    Recreate each policy guide as a project

    Create one Paxmod project per surface (public chat, private or clan chat, usernames). Each has its own API key, categories, thresholds and allow list. Rules that do not map to a fixed category go into an Advanced mode policy written in plain English.

  4. 4

    Swap the call

    Point your moderation call at POST https://www.paxmod.com/api/v1/text with an Authorization: Bearer header, and send the text as message with user_id and context_id. Act on result === "flagged", or read content_moderation.categories for per-category scores.

  5. 5

    Shadow-run and tune thresholds

    Send the same live messages to both Community Sift and Paxmod, compare decisions, and adjust each category threshold in the console until Paxmod matches or beats your current catch rate and false-positive rate.

  6. 6

    Cut over before your contract ends

    Move traffic to Paxmod in stages, then remove the Community Sift integration before the contract end date your Microsoft account team gave you.

Frequently asked questions

Is Community Sift shutting down?

Yes. Microsoft has told existing Community Sift customers that their contracts will be honored until they expire and will not be renewed. Microsoft has not announced this publicly, and its product page is still live, but the wind-down has been underway since late 2025.

When does Community Sift stop working?

There is no single public date. Microsoft is handling it customer by customer, so your real deadline is the end of your current contract. Ask your Microsoft account team for the date and plan to be fully migrated a few weeks before it.

Is Community Sift the same as Sift?

No. Community Sift is the content moderation platform Microsoft bought with Two Hat in 2021. Sift (sift.com) is an unrelated fraud prevention company.

What is the best Community Sift alternative for game chat?

If your main surface is text chat and usernames in a game, Paxmod is the closest fit you can start today: a real-time chat moderation API built for Unity, Roblox, Minecraft and Discord, with evasion-resistant text normalization, per-surface policies, PII and grooming-risk detection, and published pricing from $29/month. If you mostly need voice moderation, look at GGWP or Modulate. If you need video, look at Hive or Lasso.

Can Paxmod replace Community Sift's User Reputation system?

Partly, and differently. Send a user_id with every message and Paxmod keeps each player's flag history and builds a per-user risk summary in User Analysis. Studio adds conversation context, so a message is judged against the player's recent messages instead of on its own. Paxmod does not move players between Trusted and Not-Trusted states for you; your game decides the action, using those signals.

Does Paxmod moderate usernames?

Yes. Send the username to the same text endpoint. Most teams use a separate Paxmod project for usernames with stricter thresholds, the same way Community Sift let you set a stricter policy guide for names than for chat.

How much does Paxmod cost compared with Community Sift?

Community Sift has no public price; it was quoted by Microsoft sales. Paxmod is $29/month for 50,000 messages (Indie) or $149/month for 500,000 messages with image moderation and conversation context (Studio). Enterprise, with data residency in Canada or the EU and an SLA, starts at $1,500/month. Every plan starts with a 14-day free trial with 10,000 messages and every feature.

How long does a Community Sift migration take?

The code change is usually one function: a new URL, a Bearer header and different response fields. Most of the time goes into recreating your policy guides and running Paxmod next to Community Sift on real traffic to tune thresholds. Start that well before your contract ends.

Run Paxmod next to Community Sift this week

Start a 14-day free trial with 10,000 messages and every feature, and compare decisions on your real chat before your contract ends.

Sources

  1. Lasso Moderation: What is Community Sift? (customer notice and wind-down), and Microsoft's Community Sift product page
  2. Lasso Moderation: Community Sift alternatives (feature checklist), and Microsoft acquires Two Hat (2021)
  3. GGWP
  4. Lasso Moderation pricing
  5. Hive pricing and Hive text moderation docs

Competitor details come from public pages as of November 2026. If something here is out of date, tell us and we will fix it.

Related links