Quick Start
Create your first memory in under 5 minutes using the CLI, SDK, or API.
Get up and running with MemNexus in minutes. Choose your preferred interface below.
Prerequisites
You need a MemNexus API key. Sign up at memnexus.ai to get one. Keys follow the format:
cmk_live_<id>.<secret>Save your API key immediately — the full secret is only shown once during creation.
Create Your First Memory
Install the CLI
npm install -g @memnexus-ai/cliAuthenticate
mx auth login --interactivePaste your API key when prompted. Verify it worked:
mx auth statusCreate a memory
mx memories create \
--content "MemNexus uses Neo4j for graph storage and OpenAI embeddings for semantic search." \
--topics "memnexus,architecture,decision" \
--importance 0.8Search for it
mx memories search --query "what database does memnexus use"The semantic search finds the memory even though you searched by meaning, not exact keywords.
Install the SDK
npm install @memnexus-ai/mx-typescript-sdkCreate and search
import { MemnexusClient } from "@memnexus-ai/mx-typescript-sdk";
const client = new MemnexusClient({
apiKey: process.env.MX_API_KEY,
});
// Create a memory
const memory = await client.memories.create({
content: "MemNexus uses Neo4j for graph storage and OpenAI embeddings for semantic search.",
topics: ["memnexus", "architecture", "decision"],
importance: 0.8,
});
console.log("Created:", memory.id);
// Search by meaning
const results = await client.memories.search({
query: "what database does memnexus use",
limit: 5,
});
for (const result of results.data) {
console.log(`[${result.score.toFixed(2)}] ${result.memory.content}`);
}Create a memory
curl -X POST https://api.memnexus.ai/api/memories \
-H "Authorization: Bearer $MX_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"content": "MemNexus uses Neo4j for graph storage and OpenAI embeddings for semantic search.",
"topics": ["memnexus", "architecture", "decision"],
"importance": 0.8
}'Search for it
curl -X POST https://api.memnexus.ai/api/memories/search \
-H "Authorization: Bearer $MX_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"query": "what database does memnexus use",
"limit": 5
}'What Just Happened?
When you created that memory, MemNexus:
Generated a vector embedding of the content using OpenAI
Extracted the topics you specified
Stored the memory node and topic relationships in Neo4j
Made it immediately searchable via semantic and keyword search
When you searched, MemNexus ran a hybrid search — combining vector similarity with full-text matching — and ranked results using Reciprocal Rank Fusion (RRF).