documentation · v1
Siftfy docs
Updated May 12, 2026
Siftfy classifies arbitrary text and returns a calibrated spam probability from a single REST endpoint. Most integrations are two lines of code; the reference below covers everything beyond that.
Quickstart
Sign in, generate an API key from the dashboard, then POST text to /v1/predict:
curl -sX POST https://api.siftfy.io/v1/predict \
-H "X-API-Key: $SIFTFY_KEY" \
-H "Content-Type: application/json" \
-d '{"text": "Win a free iPhone! Click here NOW"}'
# {
# "spam_probability": 0.987,
# "likelihood": "high"
# }The response is a softmax probability between 0 and 1. We suggest a threshold of 0.85 for hard blocking; queue or soft-flag for anything between 0.5 and 0.85.
Where to next
- Authentication → how API keys work, where to put them, what 401 looks like.
- POST /v1/predict → full request and response reference, including curl, Python, and JavaScript samples.
- Errors → every status code we return and when each one fires.
- Rate limits → per-tier quotas and the headers you can monitor to stay inside them.
- SDKs → official Python client, plus what to do until other languages ship.
- Changelog → model versions, API changes, and notable releases.
Plans & pricing
Free tier covers 10,000 requests per month at no cost; Pro is $9 plus $0.0001 per request after the included 1M. See pricing for the full breakdown, and rate limits for per-minute caps.