LinkedIn Viral Post Predictor (with account)
Betathink-reason-learn/linkedin-viral-post-predictor-with-accountScores a LinkedIn post against the author's own account and audience, and shows what the account changes.
Like linkedin-viral-post-predictor, but reads the post against the author's LinkedIn account: followers and how their recent posts performed. Much sharper (0.90 viral AUC on unseen authors; its top 2% go viral 17% of the time).
The first call with linkedin_url reads the account (10-90 s) and returns a signed account_snapshot. Send the snapshot back on later calls to skip the lookup: those cost the same as a plain score. Send both and an expired snapshot (7 days) is refreshed automatically.
Quality
- Viral AUC, unseen authors
- 0.90
- Viral rate in top 2%
- 17%
- vs 1% at random
Limits
Where the model was trained, and where to trust it less.
- Trained on posts by authors with 1k-20k followers about startups, AI and venture capital, July-October 2026. Outside that range the response carries a warning.
- Pulled up / pulled down are what the model noticed, not advice.
Input
| Field | Type | Description |
|---|---|---|
| textrequired | text | Post. Up to 3,000 characters. |
| media | select | Attachment. textimageimagesvideodocument |
| image | image | Image or video frame. |
| quote | checkbox | Quotes another post. |
| linkedin_url | url | LinkedIn profile. |
| account_snapshot | text | Account snapshot. From an earlier response. Skips the account lookup. |
Output
The output object of POST /v1/predict. Over chat completions, the same JSON is the reply content.
| Field | Type | Description |
|---|---|---|
| level | string | bottom half | 50-75th | 75-90th | 90-98th | top 2% |
| chance_hit | number | Probability of a hit: top 5% of posts by reactions among authors with a similar following. |
| chance_viral | number | Probability of going viral: top 1%. |
| base_chance_hit, base_chance_viral | number | The base rates, for comparison. |
| content_score | number | 0-1 score of the content alone. |
| pulled_up, pulled_down | array | Up to 5 checks each: {check, answer, effect}. What the model noticed, not advice. |
| image_reading | string | null | How the image was read. Send it back as image_reading to rescore edits without the image. |
| warnings | string[] | E.g. the author is outside the training range. |
Call it
POST /v1/predict with your API key, or any OpenAI SDK pointed at our base URL.
curl https://api.thinkreasonlearn.com/v1/predict \
-H "Authorization: Bearer $TRL_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "think-reason-learn/linkedin-viral-post-predictor-with-account",
"input": {
"text": "Your text here"
}
}'