Working with a team? The same memory graph works across a whole team, with roles, agentic review on shared writes, and an audit trail. See Teams
Memory API

Add long-term memory to your LLM stack

A simple API for storing and retrieving structured memory. Give your product the context it keeps losing, across sessions and deployments.

Start building
Important distinction

This is not a vector database

Vector search gives you similarity. Maindex gives you memory.

  • Structured memory, not flat embeddings
  • Linked concepts, not isolated chunks
  • Persistent knowledge, not disposable context
Start building with memory
Why this matters

Without memory, your product resets every session

Try the API
POST /v1/memories
{
  "headline": "User prefers annual pricing",
  "body": "Discussed in onboarding call. Strong preference for annual billing with upfront discount.",
  "tags": ["pricing", "preferences", "onboarding"]
}

Stored, tagged, linked, and queryable. That's it.

What you get

Structured storage

Memories with tags, links, and collections, not raw embeddings.

Semantic + relational recall

Query by meaning and follow connections between concepts.

REST API + MCP

REST for your backend. MCP for direct agent integration. Use both.

Start building with memory

14-day trial. Card required. Cancel anytime. Add persistent context to your product today.

Start building