Hlas-1 is a 3.5B open-weight text-to-speech model that whispers, laughs, cries, and follows custom emotion and pacing instructions. Shape delivery with inline tags, clone a voice from a short reference clip, and generate multilingual speech. Built as a spare-time project on Breeze TTS 2, with 24 kHz audio and a streaming runtime. Free for research and non-commercial use.
Fous turns websites without APIs into reusable data APIs. Paste a URL and describe what you need. Fous builds and verifies an API you can call whenever you need it. Find matching APIs using natural language, call typed operations directly, or replay saved requests for fresh data without repeating discovery. Public and private APIs share one key, centralized billing, and per-request budgets.
Fous turns photos into location context. Upload an image and Fous estimates where it was taken by analyzing visible scene details, without relying on EXIF, GPS, or embedded metadata. The first preview focuses on San Francisco, with confidence-scored results, map review, and a workflow built for verification, research, investigations, trust and safety, and personal photo archives.
