INTEGRATION GUIDE · BETA · SOURCE CHECKED 16 NOV 2026

Migrate Coral Talk from the Perspective API (before 31 Dec 2026)

Perspective-shaped requests, independent scores. Change the connection, then validate the moderation decisions.

Google’s Perspective API notice gives 31 December 2026 as the shutdown date. Plan and test your migration before then.

1. Prepare a staging copy

Save the integration version, settings, current thresholds and old scores before changing providers. Create a Decision Models key in the dashboard; store it only on your server. Examples use environment variables, never real keys.

Coral request builder and URL construction shows comment analysis and moderation feedback calls. The configured model becomes the requested attribute; the endpoint is configurable in the Coral custom endpoint and key fields.

2. Change the endpoint and key

In Coral’s Configure → Moderation → Toxic comment filter, set Custom endpoint to the URL below, including /v1alpha1. Replace the API key with your Decision Models key. Keep your chosen model, then validate it. The sender appends /comments:analyze or /comments:suggestscore, so do not paste the full analyze URL.

integrations.perspective.endpoint = "https://moderation.decisionmodels.io/v1alpha1"
integrations.perspective.key = <your Decision Models key>
integrations.perspective.model = "TOXICITY"

Disable “Allow Coral to send moderation actions to Google” (integrations.perspective.sendFeedback) if you do not need an acknowledgment-only feedback path. Its label still refers to Google even after changing providers.

Attributes and compatibility

The default model is TOXICITY, a built-in attribute (Coral defaults). Coral can request a different configured model: audit your own value against the list below. Its locale mapping sends en, es, fr, de or pt explicitly; these valid hints are accepted and echoed. Untested codes appear in X-DM-Language-Untested (English: 5 attributes; German: offensive-language proxy only). Test quality for the language your tenant actually sends.

Built-in and experimental attribute names

Built-in: TOXICITY, IDENTITY_ATTACK, INSULT, PROFANITY, THREAT. Built-in does not mean score parity with Perspective.

Experimental, not evaluated and uncalibrated: SEVERE_TOXICITY, SEXUALLY_EXPLICIT, FLIRTATION, SPAM, ATTACK_ON_AUTHOR, ATTACK_ON_COMMENTER, INCOHERENT, INFLAMMATORY, LIKELY_TO_REJECT, OBSCENE, UNSUBSTANTIAL. Responses identify requested experimental attributes in X-DM-Experimental-Attributes. Unknown attributes return 400. Check the current attribute and language documentation before enabling automatic actions.

Re-tune the thresholds

Coral defaults sets the default to 80% (0.8). The admin UI explains that comments above the threshold go to the pending queue. Re-tune before restoring automatic queue decisions.

Use the migration tool with representative comments, their saved Perspective scores and optional human labels. Compare flag rates, flagged-set overlap and, where labels exist, precision and recall. Matching a flag rate does not guarantee the same flagged comments or accuracy. Tune each attribute and action separately; do not copy the old numbers into production without review.

3. Test and re-tune

With DM_MODERATION_KEY set privately, this smoke command should return HTTP 200 and a numeric attributeScores.TOXICITY.summaryScore.value between 0 and 1. It checks transport and response shape, not moderation quality.

curl --fail-with-body --max-time 15 \
  'https://moderation.decisionmodels.io/v1alpha1/comments:analyze' \
  -H "Authorization: Bearer $DM_MODERATION_KEY" \
  -H 'Content-Type: application/json' \
  --data '{"comment":{"text":"Thank you for explaining."},"requestedAttributes":{"TOXICITY":{}},"doNotStore":true}'

After the smoke request, submit staged comments through Coral and confirm both the user warning and pending queue behavior. If feedback remains enabled, approve and delete a staged comment and inspect the feedback response; a successful response does not mean the model learned from it.

In staging, test benign, abusive, borderline, quoted and multilingual comments; compare actual queue or warning actions with human decisions. Also test invalid keys, untested-language warnings, 429 rate limits and a timeout. Keep a manual moderation fallback rather than interpreting a failed request as a safe score. Check limits and errors and plan quotas and request rates.

4. Roll out with a fallback

Coral’s feedback expectations and locale mapping need special attention. The replacement acknowledges comments:suggestscore without storing feedback. Retain your moderator workflow, and check application logs without exposing key query parameters.

Keep the saved version and settings so you can reverse this change. Before Google’s shutdown, you can restore the Google endpoint together with its Google key and old thresholds if that service remains available to your account. After shutdown, use a tested alternative or manual review. Never send a Decision Models key to the Google host, or a Google key to Decision Models.

Shared gaps: scores come from an independent model; sentence-level span annotations are unavailable (one whole-text span only). Only PLAIN_TEXT is accepted. Valid explicit language hints are accepted and echoed. Untested hints or detected languages carry X-DM-Language-Untested. Tested languages: English (5 attributes) and German (offensive-language proxy only). This is not evidence of language quality. Suggested-score feedback is acknowledged but not stored or used to train the model. See the full API contract.

Source notes

These instructions follow the linked source revisions, inspected on 16 November 2026. Your installed release may differ. Patches and examples are migration recipes, not upstream endorsements or a claim of an end-to-end verified integration.

Coral request builder and URL construction · Coral defaults · Coral custom endpoint and key fields