What we've built.

Academic timetabling is an NP-hard constraint satisfaction problem. Faculty availability, room capacity, course dependencies, lab requirements, and administrative constraints all have to resolve simultaneously. The solution space grows exponentially with institution size.

Most colleges solve this with spreadsheets, spending 2–4 weeks per semester and still discovering conflicts after publication. Slotly replaces that with constraint optimization that generates conflict-free schedules in minutes.

Serving 10,000+ daily users across multiple institutions.

Voice cloning has been cloud-dependent, account-gated, and privacy-invasive by default. Say It Like runs entirely on-device: local ML models, no server calls, no account required. Clone any voice from an audio file or microphone recording and generate speech from text, privately, on your Mac.

It integrates into macOS at the system level via Services, so you can select text in any app and speak it in a cloned voice without switching context.

Available on the Mac App Store.

Colleges accumulate years of domain-specific knowledge: admissions policies, fee structures, course catalogs, placement data, but no way to make it instantly accessible. Generic chatbots hallucinate. Standard RAG systems lose context. And none of them understand the nuance of institutional knowledge.

Jiffy uses Integrated Context Architecture (ICA) to build a purpose-fit knowledge agent for each institution. During admission cycles, when query volumes spike to 500+ per day, Jiffy handles them accurately, in whatever language the student writes in.