BotPsy: On-Device Personal Assistant with a Sub-Billion-Parameter Agentic Language Model
3.42T1.5 sourceHugging Face Blog
Source record
Published by Hugging Face Blog (T1.5 source). The original is at https://huggingface.co/blog/qvac/botpsy.
Pipeline notes
The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.
SummaryTether AI Research presents BotPsy, a 0.8-billion-parameter model built on Qwen3.5-0.8B for on-device function calling and agentic workflows. It targets routine, low-to-medium-difficulty tasks across research, commerce, device control, documents, calendars, email, and messaging, with self-reported tool-use performance at its size.
Why it mattersProvides a concrete example of privacy-conscious task agents that operate without frontier APIs. Its capability-to-compute trade-off informs local assistant workload planning.

Cited by
No citations on record.