Do Chazz.ai characters remember previous conversations?

Yes. Facts extracted from a conversation persist between sessions, decay when unused, and consolidate as they accumulate. How far back a reply reaches is a per-conversation setting.

Does the character remember what I told someone else?

No. Memory belongs to the conversation, not the character card. The same character talking to two people keeps two separate memories.

Characters that remember · Chazz.ai

How memory works on Chazz.ai: facts are extracted from a conversation, kept between sessions, decay when unused and consolidate when they accumulate, with the depth under the reader's control and priced accordingly.

Chazz.ai characters remember across sessions. Facts are extracted from what happened rather than the whole transcript being re-sent, unused memories decay, related ones consolidate, and how far back a reply reaches is a setting the reader chooses and pays for per turn.

The problem it solves

The usual approach to giving a model a memory is to paste the transcript back in front of every turn. That works until the conversation is longer than the context window, at which point it silently starts forgetting the beginning while appearing to remember everything, and it charges you for re-reading the same text on every single turn.

What happens instead

Extraction: After a turn, what actually happened is written down as facts rather than kept as raw text. "She admitted she has been lying about the map" is a fact; the four paragraphs it happened in are not.
Decay: A fact that stops being relevant loses weight over time. Without this a character three hundred turns in is buried under trivia and reads as distracted.
Consolidation: Related facts merge into a larger one. Six separate notes about the same argument become the argument.
Retrieval: Only what bears on the current turn is put in front of the model. That is why depth is a price rather than a switch.

The control you get

Memory depth is a slider, not a mode. Shallow is cheap and quick and suits banter; deep costs a multiplier on the turn and suits a story you have been running for weeks. The multiplier is shown before you spend it. The reference is in the memory documentation.

Memory is per conversation and belongs to the conversation owner. It is encrypted at rest under a per-owner key, and it is included in an account data export.

What it cannot do

It does not read your other conversations. It does not follow a character between two readers: the same card talking to two people keeps two separate memories, which is the only answer that is not a privacy problem.

Questions

Do Chazz.ai characters remember previous conversations?: Yes. Facts extracted from a conversation persist between sessions, decay when unused, and consolidate as they accumulate. How far back a reply reaches is a per-conversation setting.
Does the character remember what I told someone else?: No. Memory belongs to the conversation, not the character card. The same character talking to two people keeps two separate memories.

Elsewhere on Chazz.ai

This document is the server-rendered form of /features/memory. The interactive page needs JavaScript, and shows the same content.

The complete documentation of this service is available as one plain-text file at /llms-full.txt. A short summary, with the operator's policy for automated readers, is at /llms.txt. Changes are published as a feed at /rss.xml.

Chazz.ai is operated by Cognitive Industries, ABN 62 794 528 747, Australia. Contact: [email protected].