VISIONA Perspective API alternative, before the shutdown.
Google is sunsetting Perspective API: it stays active until 31 December 2026, and Google will not offer direct migration support. If your comments are in Turkish or English, here is where each of its attributes goes in Detectivision, and what the code looks like before and after.
What Google has announced
From the notice on perspectiveapi.com, as read in October 2026. Detectivision is not affiliated with Google or Jigsaw, the makers of Perspective API.
Active until 31 December 2026
The service remains active until 31 December 2026, and will no longer be in service after 2026.
No more quota
Requests for usage and for higher quota were handled until February 2026.
No migration help
Google will not offer direct migration support: each team moves on its own.
Where each Perspective attribute goes
Perspective scores attributes; Detectivision raises flags, each with its confidence. Most attributes have a flag, some do not.
| Perspective | Detectivision | Note |
|---|---|---|
| TOXICITY | flagged | There is no single toxicity score: flagged is true when any flag you asked for was raised. |
| SEVERE_TOXICITY | none | No flag of its own. Draw a stricter line on the confidence of the flags raised. |
| IDENTITY_ATTACK | hate, racism, sexism | Three flags, so your rules can treat them apart. |
| INSULT | insult | Insults and personal attacks on someone. |
| PROFANITY | none | Not among the flags every account has. Swearing aimed at someone is an insult. |
| THREAT | none | No flag of its own. |
| SEXUALLY_EXPLICIT | sexual | Sexual language, and sexual harassment of someone. |
| FLIRTATION | none | No flag of its own. |
Two flags have no Perspective attribute: self_harm, for talk of suicide or self-harm, and bad_habits, for drug use, drinking, smoking or gambling talked up.
The same check, both ways
curl -X POST \
"https://commentanalyzer.googleapis.com/v1alpha1/comments:analyze?key=$PERSPECTIVE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"comment": { "text": "You are an idiot" },
"languages": ["en"],
"requestedAttributes": { "INSULT": {}, "IDENTITY_ATTACK": {} }
}'curl https://api.detectivision.ai/api/v1/moderation \
-H "Api-Key: $DETECTIVISION_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"type": "text_moderation",
"input": "You are an idiot",
"flags": ["insult", "hate", "racism", "sexism"]
}'Detectivision takes an API key in the Api-Key header, and the type of content in type. Without flags, every text flag of your account is checked.
Reading the answer
const score = body.attributeScores.INSULT.summaryScore.value
if (score > 0.8) hide(comment)const insult = body.flags.find((f) => f.flag === "insult")
if (insult && insult.confidence > 0.8) hide(comment)The two scores are not on the same scale. A flag is raised from 0.5 by default; start from flagged, then tune your line on your own comments.
What is different
Languages
Detectivision understands Turkish and English. Perspective covers more languages; if your comments are in others, keep that in mind.
Price
Perspective was free. Detectivision is paid in credits: a text costs 1 credit per started 2,500 characters by webhook, twice that at once.
Images too
The same API checks images for sexual content, violence, hate and scams.
Actions
Moderators can act on each request from the dashboard’s Logs, and your system gets every action by webhook.
Turkish
Insults, hate, racism, sexism and sexual language in Turkish comments.
Toxicity, flag by flag
Seven text flags for toxic comments, each with its confidence.
Frequently asked
When does Perspective API stop working?
Google says the service remains active until 31 December 2026 and will no longer be in service after 2026.
Will Google help us migrate?
No. Its notice says Google will not offer direct migration support.
Is there a TOXICITY score in Detectivision?
No single score. flagged tells you whether any of your flags was raised, and each flag raised has its own confidence from 0 to 1.
Which languages does Detectivision support?
Turkish and English. Text in other languages is checked, but the results are not reliable.
Can I try it on my own comments first?
Yes. Start the free trial, then send your comments through the dashboard’s playground or the API and compare the answers with your Perspective scores.