#0001
Vercel warns AI endpoints are prime targets for inference theft
70radar
BotIDVercel bot defense tool — per-request automation checks
Cheap HTTP checks do not match $2-per-prompt economics. Verify every AI request, or one abused endpoint can turn into a five-figure bill.
- Attackers wrap custom endpoints in
OpenAIorAnthropicadapters, then resell stolen inference through standard SDK clients. - Session auth and IP limits fail when proxy fleets and throwaway accounts spread calls across thousands of identities.
- Playgrounds are the highest-risk shape because users control prompts, models, and parameters. Fixed server prompts lower but do not remove risk.
BotIDis Vercel's proposed gate: run bot verification on every AI request, not just signup or session start.
Source: vercel.com/blog/protecting-against-inference-theftRead original →