Voice → Understanding
Real-time speech pipeline captures your tone and words, then hands it to IDO-1 with sentiment tags and priority markers.
Feature deep dive
Live thinking: watch IDO-1 do the work.
The glowing lines you saw on the homepage aren’t gimmicks. They track how IDOLL moves data from your sensors into IDO-1, how the model reasons, and how the reply routes back into motion and speech.
Memory graph → Reasoning
IDO-1 looks up who and what you mentioned before connecting the dots—deadlines, promises, motivations.
Motion & light → Response
As IDO-1 speaks, IDOLL mirrors the thought process with beams, lights, subtle head tilts, and the animated graph you saw earlier.
Live graph
Model promises
Why build IDO-1 ourselves?
- Fast enough to feel alive: IDO-1 targets sub-second turnarounds for short replies so the physical robot never feels laggy.
- Long-context memory: the model can reference weeks of conversation without losing the thread.
- Mix of logic + empathy: benchmarks show IDO-1 holds its own on reasoning tasks while staying tuned for emotional cues.