Sandeep Kamble

Sandeep Kamble

SensfrxSensfrx
Fighting frauds with AuthSafe

Badges

Tastemaker
Tastemaker
Veteran
Veteran
Gone streaking
Gone streaking

Maker History

Forums

30d ago

PROMPTPurify - Tiny prompt-injection firewall for LLM chat apps. 14 MB.

Every OSS prompt injection guard needs a GPU, a sidecar, or 7 GB of disk. PromptPurify is 14 MB, CPU-only, runs inside your Node process. Zero network calls. $0/call. Two layers: deterministic structural firewall + 14 MB ML classifier. Benchmarked against Meta Prompt-Guard, ProtectAI, deepset same inputs, same scoring code, reproducible on your laptop. Battle-tested by 2,458 red teamers before release. MIT. Drop in with 3 lines of code.

5yr ago

AuthSafe - Block fraud not legitimate customers

Start with identifying frauds and increase ROI.
The cognitive engine’s model gets trained from end-users, separates good users & bad users, and provides real-time solutions to detect and prevent online fraud attacks.
View more