Google's Perspective API Is Shutting Down: A Migration Guide for Forms and Comments

If you have been quietly relying on Google's Perspective API to score toxicity in the text your users submit, this is your warning: it is going away. A notice on the Perspective API homepage confirms the service is sunsetting, with the shutdown set for the end of December 2026 and new usage or quota requests accepted only until February 2026. Google's stated reasoning is blunt — AI has moved on, and there is "less demand for a standalone tool specific to this area." There is no official migration support.

For a free tool that has quietly powered comment moderation and form scoring across thousands of sites for the better part of a decade, that is a big deal. If Perspective sits anywhere in your stack, you have a deadline now.

What Perspective Actually Did (and Where People Used It)

Perspective, built by Google's Jigsaw team, took a chunk of text and returned probability scores across attributes like toxicity, insult, profanity, threat, and identity attack. It was free, it was easy to call, and it was good enough that a lot of developers wired it into places it was never really marketed for.

The obvious use case was comment sections. But the quieter, more widespread use was forms: scoring the message field on a contact form, the free-text box on a sign-up, the "tell us about your project" field on a lead form — anywhere a stranger could type something you did not want landing in your inbox or your database. Perspective let a solo developer add toxicity filtering with a single API call and no budget line.

That is exactly the setup now on a clock.

Why "Just Find Another Free Toxicity API" Is the Wrong Question

The instinct is to swap Perspective for the nearest free equivalent and move on. Resist it, for two reasons.

First, most Perspective replacements being promoted right now are framed around chat and user-generated content feeds — comment streams, community platforms, social apps. They score toxicity, and nothing else. If your real problem is form submissions, toxicity is only one of the things you care about. The junk that clogs a contact form is overwhelmingly spam and promotional garbage, not slurs. A pure toxicity API leaves your biggest category — spam — completely unaddressed.

Second, "free forever" was the trap the first time. Perspective was free, and now it is being switched off with a few months' notice and no migration help. Building your moderation on a dependency you do not pay for means you have no leverage and no SLA when the owner decides to walk away. reCAPTCHA users learned the same lesson in 2024 when Google cut the free tier from a million calls a month to ten thousand.

So the right question is not "what is the closest free clone of Perspective?" It is "what do I actually need to catch on my forms, and what will still be here in three years?"

What to Replace It With

For form submissions specifically, you want multi-category scoring, not single-attribute toxicity. A submission should be checked for:

• Spam — the promotional junk and link-dropping that makes up the bulk of form abuse
• Profanity — the obvious stuff, before it reaches a human reviewer
• Hate speech — the reputational and legal-liability category Perspective was actually used for
• Junk / low-quality content — gibberish, form-filler, and keyboard-mashing bots

Babelshield was built for exactly this shape of problem: a single API call returns scores across all four categories, with thresholds you configure per form. That is the practical difference from a toxicity-only tool — you replace the one thing Perspective did and cover the spam problem that Perspective never touched, in the same integration.

The Migration, Step by Step

Moving off Perspective is not dramatic, but do it deliberately rather than in a panic in December.

1. Find every call. Grep your codebase and your no-code integrations for commentanalyzer.googleapis.com and analyzeComment. Do not forget serverless functions, Zapier/Make scenarios, and anything a contractor wired up two years ago. List every form and surface that depends on it.

2. Map your attributes to categories. For each call, note which Perspective attributes you were reading (TOXICITY, PROFANITY, INSULT, and so on) and what threshold you were acting on. This tells you which Babelshield categories each form needs and roughly where to set the thresholds.

3. Swap the call. A Perspective request that looked like this:

// Old: Perspective API
const res = await fetch(
  `https://commentanalyzer.googleapis.com/v1alpha1/comments:analyze?key=${KEY}`,
  {
    method: "POST",
    body: JSON.stringify({
      comment: { text: submission },
      requestedAttributes: { TOXICITY: {}, PROFANITY: {} }
    })
  }
);

becomes a single multi-category call:

// New: Babelshield
const res = await fetch("https://api.babelshield.com/v1/score", {
  method: "POST",
  headers: { "Authorization": `Bearer ${API_KEY}`, "Content-Type": "application/json" },
  body: JSON.stringify({ text: submission })
});
const scores = await res.json();
// { spam: 0.94, profanity: 0.02, hate: 0.00, junk: 0.88 }
if (scores.spam > 0.8 || scores.hate > 0.6) reject();

(Endpoints and field names are illustrative — check the current Babelshield docs for exact values and your keys.)

4. Run both in parallel first. For a week or two, call the new service alongside Perspective and log both results without acting on the new one. This gives you real data on where your thresholds should sit and catches any surprises before they affect live traffic.

5. Cut over and delete the old keys. Once your thresholds are dialled in, flip the switch, then revoke the Perspective keys so nothing silently keeps calling a service that is about to disappear.

Do Not Wait Until December

The temptation with a year-end deadline is to leave it for later. Two reasons not to. New Perspective quota requests close in February 2026, so if your usage grows you may hit a wall long before the shutdown. And a rushed migration in the last week of December — over the holidays, with reduced staff — is exactly how a form silently stops being moderated and nobody notices until the spam and abuse are already in the CRM.

Perspective was a good free tool while it lasted. Its retirement is a nudge to put form and comment moderation on a footing you actually control — one that covers spam as well as toxicity, and that will not vanish on you the next time priorities shift at a large platform.

References

Perspective API. Sunset notice, perspectiveapi.com (accessed 2026).

Google Jigsaw. Perspective API documentation and attribute reference.

Google Cloud. reCAPTCHA billing changes, 2024.