Memory icons

The 22 memory icons in Agents, described

Agent bookmarkagent-bookmark

An agent beside a bookmark — a place an agent marked to return to.

Agent contextagent-context

An agent beside a pair of brackets — the context window an agent carries into each step.

Episodic memoryepisodic-memory

Notes beside a clock — what an agent remembers happening, and when.

Long-term memorylong-term

Long-term memory — what survives the session, persistent and durable.

Memorymemory

Agent memory — what an AI agent writes down and reads back between steps and sessions.

Memory clearmemory-clear

Memory clear — forget it all, reset and wipe an agent's memory.

Memory consolidatememory-consolidate

Several notes folding into one compact memory — consolidating what an agent has learned.

Episodic memorymemory-episodic

Episodic memory — what happened, in order, an agent's record of events.

Memory loadmemory-load

An arrow leaving stored notes — an agent loading something from memory.

Long-term memorymemory-long-term

Long-term memory — what an agent keeps between runs and can recall later.

Memory pinmemory-pin

Memory pin — this memory is not allowed to fade, pinned to be kept.

Memory readmemory-read

Memory read — fetch what an agent stored earlier and bring it back into context.

Memory savememory-save

An arrow writing into stored notes — an agent saving something to memory.

Memory searchmemory-search

A magnifying glass over stored notes — searching an agent's memory.

Semantic memorymemory-semantic

Semantic memory — what things mean and how facts join to each other.

Short-term memorymemory-short-term

Short-term memory — what an agent is holding right now, within the current session.

Memory writememory-write

Memory write — put something away in an agent's memory for later.

Procedural memoryprocedural-memory

Notes beside a play button — the procedures an agent remembers how to run.

Scratch filescratch-file

A page with a scribble across it — the scratch file an agent thinks in before it answers.

Semantic memorysemantic-memory

Notes beside a model core — the facts and concepts an agent keeps as memory.

Statestate

State — what is true right now, a snapshot of the current values.

Working memoryworking-memory

Working memory — what a model or agent is holding while it thinks.

In code, each is one import — import { AgentBookmark } from "@iconmind/react/icons/agent-bookmark" — and the same name in Vue, Svelte, Solid, Preact, React Native, Astro, Blade and Flutter.

Narrow it by tag