mirror of
https://github.com/supermemoryai/supermemory.git
synced 2026-10-10 03:28:14 +00:00
5437f9743c
The Overview didn't carry the positioning the product earns. Lead with proof and the technical edge instead of generic capability cards: - Open with the category + the #1-on-every-benchmark claim and the "reasons over a directed knowledge graph" thesis - Add a benchmarks table (LongMemEval 81.6%, LoCoMo, ConvoMem — all #1) - Replace the generic "what you can build" grid with "What makes Supermemory different": reasons (not retrieves), directed graph, whole context stack in one system, multi-modal, built to build on, run anywhere - Keep deployment callout, how-it-works, the three context approaches, and the start-building / ways-to-use cards Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
115 lines
6.3 KiB
Text
115 lines
6.3 KiB
Text
---
|
|
title: "Overview — What is Supermemory?"
|
|
sidebarTitle: "Overview"
|
|
icon: "book-open"
|
|
---
|
|
|
|
Supermemory is the memory and context engine for AI agents — and it's the state of the art, ranked **#1 on every major memory benchmark**: [LongMemEval](https://github.com/xiaowu0162/LongMemEval), [LoCoMo](https://github.com/snap-research/locomo), and [ConvoMem](https://github.com/Salesforce/ConvoMem).
|
|
|
|
Your AI forgets everything between conversations. Supermemory fixes that. It learns from every interaction, **reasons over a directed knowledge graph** to resolve contradictions and track how facts change over time, forgets what's expired, and serves the right context at the right moment — through one API.
|
|
|
|
## State of the art, by the numbers
|
|
|
|
Most memory systems claim to be better. Supermemory is measurably ahead — first place across all three independent benchmarks the field uses to grade AI memory.
|
|
|
|
| Benchmark | What it measures | Supermemory |
|
|
|---|---|---|
|
|
| [LongMemEval](https://github.com/xiaowu0162/LongMemEval) | Long-term memory across sessions, with knowledge updates | **81.6% — #1** |
|
|
| [LoCoMo](https://github.com/snap-research/locomo) | Fact recall across long conversations (multi-hop, temporal, adversarial) | **#1** |
|
|
| [ConvoMem](https://github.com/Salesforce/ConvoMem) | Personalization and preference learning | **#1** |
|
|
|
|
Read the full methodology on the [research page](https://supermemory.ai/research), or reproduce it yourself with [MemoryBench](/memorybench/overview), our open-source benchmarking framework.
|
|
|
|
## What makes Supermemory different
|
|
|
|
Most "memory layers" are a vector store that retrieves the nearest chunk. Supermemory is an engine that *understands* — which is why it tops the benchmarks instead of just claiming to.
|
|
|
|
<CardGroup cols={2}>
|
|
<Card title="It reasons — not just retrieves" icon="brain" href="/concepts/how-it-works">
|
|
Resolves contradictions, tracks temporal change, follows multi-hop relationships, and forgets expired facts automatically. Reasoning is why it wins every benchmark.
|
|
</Card>
|
|
<Card title="A directed knowledge graph" icon="git-fork" href="/concepts/graph-memory">
|
|
One evolving directed graph and a single ontology across all your data — not a flat vector store, and not a raw graph you have to traverse yourself.
|
|
</Card>
|
|
<Card title="The whole context stack, one system" icon="layers" href="/concepts/super-rag">
|
|
Memory, user profiles, hybrid search (RAG), connectors, and file processing — together, sharing one context pool. No stitching five tools together.
|
|
</Card>
|
|
<Card title="Multi-modal — everything in" icon="file-text" href="/concepts/content-types">
|
|
Text, conversations, PDFs, images (OCR), video (transcription), and code (AST-aware chunking). Upload it and it just works.
|
|
</Card>
|
|
<Card title="Built to build on" icon="blocks" href="/integrations/supermemory-sdk">
|
|
One API, plus SDKs, a CLI, a memory filesystem, and MCP. Infrastructure you ship products on — not a closed box.
|
|
</Card>
|
|
<Card title="Run it anywhere" icon="server" href="/self-hosting/overview">
|
|
Hosted for zero-ops scale, or the full engine as one self-hosted binary — fully offline if you want, same API either way.
|
|
</Card>
|
|
</CardGroup>
|
|
|
|
<Tip>
|
|
**Hosted or self-hosted — same API.** Use the [managed platform](https://console.supermemory.ai) for zero-ops scale, or [run the entire engine on your own machine](/self-hosting/overview) — one binary, zero config, fully offline. Move between them by changing a single `baseURL`.
|
|
</Tip>
|
|
|
|
## How does it work? (at a glance)
|
|
|
|
|
|
|
|
- You send Supermemory text, files, and chats.
|
|
- It [indexes them intelligently](/concepts/how-it-works) and builds a directed knowledge graph on top of an entity (a user, document, project, or organization).
|
|
- At query time, it fetches only the most relevant context and passes it to your models.
|
|
|
|
## Three ways to add context
|
|
|
|
Memory, profiles, and search all draw from the **same context pool** for a given user (`containerTag`) — so they reinforce each other instead of living in silos. Mix and match as your use case needs.
|
|
|
|
#### Memory API — learned user context
|
|
|
|
|
|
|
|
Supermemory learns and builds memory for each user — extracted facts that:
|
|
- [Evolve on top of existing context about the user](/concepts/graph-memory), **in real time**
|
|
- Handle **knowledge updates, temporal changes, and contradictions**
|
|
- Power a **user profile** that acts as the default context provider for the LLM
|
|
|
|
#### User profiles
|
|
|
|
The evolving context produces a [**User Profile**](/concepts/user-profiles) — the facts your agent should **always** know, in one ~50ms call:
|
|
|
|
- **Static:** stable facts the agent should always know.
|
|
- **Dynamic:** episodic context from the last few conversations.
|
|
|
|
#### RAG — advanced semantic search
|
|
|
|
Run hybrid [search](/search) over the raw content too: advanced metadata filtering, contextual chunking, and reranking — tightly integrated with the memory engine, in a single query.
|
|
|
|
## Start building
|
|
|
|
<CardGroup cols={3}>
|
|
<Card title="Quickstart" icon="play" href="/quickstart">
|
|
Make your first API call in minutes.
|
|
</Card>
|
|
<Card title="Add your first memory" icon="plus" href="/add-memories">
|
|
Ingest text, files, and conversations into a container.
|
|
</Card>
|
|
<Card title="Search it" icon="search" href="/search">
|
|
Retrieve the most relevant context with hybrid semantic search.
|
|
</Card>
|
|
</CardGroup>
|
|
|
|
## Ways to use Supermemory
|
|
|
|
The API is the core — but you don't have to talk to it directly. Reach Supermemory however fits your workflow:
|
|
|
|
<CardGroup cols={2}>
|
|
<Card title="SDKs" icon="code" href="/integrations/supermemory-sdk">
|
|
Official TypeScript and Python SDKs, plus drop-in plugins for the AI SDK, OpenAI, LangChain, and more.
|
|
</Card>
|
|
<Card title="Command line (CLI)" icon="square-terminal" href="/cli/overview">
|
|
Manage memories, search, and scripting from your terminal — it's all `npx supermemory`.
|
|
</Card>
|
|
<Card title="Memory filesystem (SMFS)" icon="database" href="/smfs/overview">
|
|
Mount a container as a real directory your agent can `ls`, `cat`, and semantically `grep`.
|
|
</Card>
|
|
<Card title="Self-host it" icon="server" href="/self-hosting/overview">
|
|
Run the full memory engine on your own machine — one binary, zero config, fully offline.
|
|
</Card>
|
|
</CardGroup>
|