Smart idea but the concern can be that in the future, tokenization techniques and libraries may change. And also this looks like a very edge optimization to me. But overall, it deserve to exist. Good job.
Floyd is an open-source booking kernel for AI agents: TTL holds -> confirm, explicit 409 conflicts, and idempotency keys to prevent double-bookings under concurrency.
Working on Floyd (https://floyd.run) — an AI agent/workflow runtime for bringing LLM agents to production. Still shaping the MVP and figuring out where it adds the most value. Would love feedback from anyone building or deploying agents, or working with n8n/Make.
I agree with you. People using this kind of service likely already have some self-awareness about their social media habits. In the end, self-discipline and personal boundaries might be a better solution than relying on external tools. But I think we can also consider this as another approach to finding a better solution.