harnyGlossary
Memory decay
Memory decay is the gradual fading of stored memory as it gets older, so an AI agent recalls recent information more readily than old information. Memory that decays can lose details that still matter.
What memory decay is
In AI agents, memory decay means older stored facts become less likely to be recalled or used. The effect is usually modeled by age, recency or how often an item is accessed.
Decay is a design choice. Some systems let memory fade, while others keep every value until someone removes it.
Why it matters
- Fading can hide facts that are old but still true.
- Keeping everything can surface outdated values, so the agent needs a way to tell old from current.
How it works in Harny
- Memory in Harny is dated. A changed value with a newer timestamp becomes the current value, and the earlier value stays answerable by date.
- You delete memory through the Memory API:
DELETE /v1/memory/data-points/{sourceId}removes a data point and the memories built on it. Deletes are free.
Does Harny forget old memories automatically?
Memory in Harny is dated. A changed value with a newer timestamp becomes the current value, and the earlier value stays answerable by date. You remove memory by calling the delete routes in the Memory API.
Keep reading
Go on. Get Harny.
The first 500 monthly-active end users are free for good. Create an app key and make your first call in minutes.