I would like to request a persistent custom assistant feature in Kimi, similar in concept to Custom GPTs (ChatGPT) and Gems (Google Gemini).
The problem
For long-term learning, research, or professional workflows, users currently need to restate role constraints and system instructions in every new session. This causes:
- inconsistent assistant behavior,
- repeated prompting overhead,
- difficulty maintaining disciplined, multi-session workflows.
What I’m requesting
Support for custom assistant profiles that persist across sessions, allowing users to define:
- persistent system-level instructions (behavior / policy),
- a stable role (e.g., strict tutor, reviewer, research assistant),
- optional attached documents or reference knowledge,
- consistent behavior across sessions.
Why this matters for Kimi
Kimi excels at long-context understanding and analytical depth. Persistent assistants would:
- fully leverage these strengths,
- enable structured study and research over time,
- improve output consistency and reliability.
This is not model fine-tuning. It is a product/platform capability that adds a persistence and configuration layer around the model.
Thank you for considering this request.