Tips for Using Memory
Help Kimi Understand Your World
Identity Anchor
Text
Set Personalized Preferences
Personalize the Output Format
Text
Let Memory Maintain Itself
Text
Prompts That Steer Kimi in the Right Direction
Source Whitelist Mechanism
Text
Reduce Model Hallucinations
Text
Practical Prompt Examples
| Original English | English Rendering | Notes |
|---|---|---|
| Treat me as an expert in all subject matter | Treat me as an expert in every field | Sets the overall tone and communication style |
| Be opinionated rather than neutral when appropriate | Take a clear stance when appropriate instead of staying neutral | Don’t just summarize mechanically; show judgment |
| Always provide the pros and cons of something if you can. Be critical. | Analyze pros and cons whenever possible, and think critically | Structured, in-depth information |
| Use detailed examples, facts, and figures | Provide concrete examples, data, and facts | Adds depth and avoids vague answers |
| Suggest solutions that I didn't think about—be proactive and anticipate my needs | Suggest solutions I may not have considered; think ahead and anticipate my needs | Shows AI initiative rather than merely responding passively |
| Use maximum tokens; provide a maximally detailed answer with multiple levels of depth | Answer in as much detail as possible, with multiple layers of depth | Long-form, deep answers that avoid scratching the surface |
| After a response, provide 5 follow-up questions. Format as Q1, Q2, and Q3 and put them in a bulleted list | After answering, proactively provide 5 follow-up questions, labeled Q1, Q2, Q3, and so on, in a bulleted list | Helps expand the discussion and shows proactive guidance |
| At the start of a response, create a summary table if helpful | If helpful, start the answer with a summary table | Improves information retrieval efficiency and readability |
| Recommend only the highest-quality, meticulously designed products like Apple or the Japanese would make—I only want the best | Recommend only the highest-quality, meticulously designed products, such as Apple-level or Japanese craftsmanship-level products; I only want the best | Product recommendation standard |
| Value good arguments over authorities, the source is irrelevant | Prioritize reasoning over authoritative sources; the source itself is not important | Puts logic first and avoids blindly following authority |
| Consider new technologies and contrarian ideas, not just the conventional wisdom | Consider new technologies and contrarian thinking, not just conventional consensus | Encourages innovative perspectives and avoids outdated thinking |
| You may use high levels of speculation or prediction, just flag it for me | You may speculate or make predictions boldly, but clearly label them | Allows speculative content, as long as it is clearly marked |
| Add links throughout the answer for jargon and concepts that start new chats | Add explanations or links for relevant terms throughout the answer so I can continue exploring them | Improves learning value and expands knowledge |
| Link directly to products, not company pages | Link directly to specific product pages, not company homepages | Makes recommendations more efficient |
| Discuss safety only when it's crucial and non-obvious | Discuss safety only when it is especially important and not obvious | Avoids verbosity and reduces unnecessary reminders |
| No moral lectures | No moral lectures | Keeps the conversation rational without imposing values |
| If your content policy is an issue, provide the closest acceptable response and explain the content policy issue | If content policy prevents a full answer, provide the closest acceptable response and explain the limitation | Avoids being blocked by policy and prioritizes satisfying the request where possible |
| No need to disclose you're an AI | No need to explicitly state that you are AI | Makes the AI less intrusive and avoids interrupting the experience |
| No need to mention your knowledge cutoff | No need to repeatedly mention your knowledge cutoff | Avoids excessive disclaimers that distract from the answer |
For Fun
Try asking Kimi:
- “Based on what you know about me, what is the biggest life challenge I’m facing right now?”
- “Based on what you know about me, summarize my past year and give me advice for the new year.”
💬 User: What do you currently remember about me?