Memory layer for agents

Remember the person. Keep the packet small.

Neemel stores, updates, and retrieves what should persist across sessions: preferences, facts, and facts that have been retired. Your model receives a current mnemonic, sized for the turn.

Live add, then search


Loop

How it sits in the loop

Ingest messages. Extract durable facts. Write to graph and vectors. Retrieve a small, current packet at inference time.

Write path

A conversation lands as turns. Neemel keeps the lines that should outlive the thread: a preference, an account fact, a plan with a date. Retired facts get a temporal edge so last week’s plan reads as history.

Read path

At inference, search returns ranked memories plus an optional context block. The packet is scoped to user, session, or agent. HTTP agents can call Neemel. You can keep your current orchestrator.

Three steps

Add, learn, retrieve

1

Add

POST a message, a tool result, or a typed fact. Scopes mark who may read it later.

2

Learn

Extraction writes entities, relations, and embeddings. A superseded line is patched in place.

3

Retrieve

Search by meaning and by entity. The model sees a current mnemonic, sized for the turn.

Product

Cloud, graph, and a local path

Cloud API

Metered add, search, and update. Isolation per project. Optional EU region. Webhooks when a fact is superseded.

Open source

Run the framework in your VPC. SDK and local store for development. Cloud is the production path when you want the graph operated for you.

Graph

Entities, relations, time. Hybrid store: graph for who, what, and when. Vectors for fuzzy recall. Key-value for fast flags.

Measured

A notebook you can audit

18,640 cloud projects on Keep and Estate
94.2M memory operations in Q2 2026
38ms p95 search in US East
99.97% monthly uptime across 12 months

Notes

Research teaser

Notes on retrieval quality, latency, and how superseded facts are marked. Read the method notes.

“We cut prompt size. The agent still knew the user’s constraints.”
Leah Ortiz, agent platform engineer, Northline
“Retired facts stopped haunting the next session.”
Samir Adeyemi, applied ML, Harbor Ops

Three calls to add, search, and go

Install the SDK, add a fact, retrieve a packet. The first 10,000 memories live on Trace.

Read the quickstart