Comparisons
Maintain your companion's story with Candy AI and Nomi AI
Work out which facts belong to the companion, which belong to a shared scene and which you want to revise. We provide Premium Plus personas, memory slots and separate chat branches. Nomi uses Backstory+ and Mind Maps for companion context, with its own group controls. Choose the information-management task before matching the feature names.
On this page
Separate background notes from the scene now underway
Organize Nomi's companion and shared-world details through Backstory+, including Backstory, Current Roleplay, Preferences and Boundaries. Put a continuing relationship fact in the appropriate background note and the immediate event in the roleplay context rather than blending them into one instruction.
Review Nomi's editable Mind Map summaries, which can update automatically and allow individual entries to be locked. Changing that summary does not delete the experiences it represents. Our slots also manage information without guaranteeing complete recall or supplying a numerical conversion to Nomi's system.
Match the change to the relevant context or media tool
| What you want to manage | Candy AI | Nomi AI |
|---|---|---|
| Establish your place in the story | Choose an available user persona for the interaction | Write the relevant shared-world details in Backstory+ |
| Change how the scene continues | Copy an eligible chat and develop an independent continuation | Guide the situation through current-roleplay and shared-context notes |
| Coordinate a cast's replies | Use memory slots for their purpose, not as equivalent speaker selection | Pick the replying companion or use automatic turn-taking |
| Give the scene a visual form | Select the companion and write the picture's visible details | Use a contextual selfie request based on the companion and ongoing chat |
Choose between revising a saved note, creating another conversation branch and coordinating a group. Even when the goal sounds similar, those operations do not necessarily retain or share the same information.
Set up the selfie scene and choose the voice arrangement separately
To request a Nomi selfie, describe the outfit, activity and place in the conversation before using the camera beside the chat field. That selfie workflow draws chiefly on the character and conversational context, rather than treating the request as a conventional standalone picture prompt.
For Nomi's paid calls, allow microphone access and account for the text thread's history being shared with the call. Its in-house voices are included, but an optional ElevenLabs voice remains under that external plan's limits. On our account, read the token costs for supported calls and voice-message listening independently.
Distinguish additional slots from messages and media spending
Use Nomi's profile menu and Credits tab to inspect purchases of extra companion slots or group rooms beyond the paid plan's included allocation. That purchase adds slot capacity, not a quantity of messages. Check which extra you want before spending the balance.
When choosing our media, read both the membership access and the charge on the specific action. A paid plan or an editable context note does not establish an unlimited allowance for all the other tools around a conversation.
Your questions
Does changing a Nomi Mind Map entry erase the experience behind it?
Treat the entry as a summary, not a direct replacement for underlying memories. Revise the relevant context with that distinction in mind; an edited summary is not the same operation as restarting the companion.
Where should I put a fact that all Nomi group participants need?
Use the group's Backstory for information intended for the whole cast. Do not rely on individual Backstory+ notes being completely shared among participants; their personal context is not automatically available in full to everyone else.
How should I describe a scene before requesting a Nomi selfie?
Establish what the character is wearing, doing and where they are in chat, then select the selfie camera. With our dedicated image tool, you instead write the visible scene into the image request; these are different ways of supplying context.