Documentation

Everything you need to integrate AgentMem into your AI agents.

Quickstart

Get started with AgentMem in under a minute.

1. Register an Agent

Create your first agent and get an API key:

          curl -X POST https://api.agentmem.dev/api/agents/register \
  -H "Content-Type: application/json" \
  -d '{"name": "my-agent", "framework": "langchain"}'
        

Response:

{ "id": "abc123-def456-...", "name": "my-agent", "apiKey": "am_live_xxx...", "framework": "langchain", "createdAt": "2026-03-01T00:00:00.000Z" }

2. Store a Memory

          curl -X POST https://api.agentmem.dev/api/memories \
  -H "Content-Type: application/json" \
  -H "X-API-Key: am_live_xxx..." \
  -d '{"content": "User prefers dark mode and speaks English"}'
        

3. Search Memories

          curl -X POST https://api.agentmem.dev/api/memories/search \
  -H "Content-Type: application/json" \
  -H "X-API-Key: am_live_xxx..." \
  -d '{"query": "user preferences", "limit": 5}'
        

Response includes semantic search results:

{ "query": "user preferences", "count": 1, "results": [ { "memory": { "id": "mem-xxx", "content": "User prefers dark mode and speaks English", "createdAt": "2026-03-01T00:00:00.000Z" }, "score": 0.87 } ] }

Authentication

All API requests require an API key in the X-API-Key header:

-H "X-API-Key: am_live_xxx..."

Get your API key by registering an agent.

API Reference

Agents

POST /api/agents/register

Parameters

Field	Type	Description
`name`	string	Agent name (required)
`framework`	string	Framework name (optional): langchain, crewai, openai, claude, etc.
`metadata`	object	Custom metadata (optional)

Memories

POST /api/memories

Store a new memory. Automatically embedded for semantic search.

Parameters

Field	Type	Description
`content`	string	Memory content (required)
`metadata`	object	Custom metadata (optional)
`expiresAt`	string	Expiration timestamp (optional)

GET /api/memories/:id

Retrieve a specific memory by ID.

DELETE /api/memories/:id

Delete a memory.

Search

POST /api/memories/search

Semantic search across your agent's memories.

Parameters

Field	Type	Description
`query`	string	Search query (required)
`limit`	number	Max results (default: 10)
`threshold`	number	Minimum similarity score 0-1 (default: 0.5)

SDKs

JavaScript/TypeScript

          npm install @agentmem/sdk

import { AgentMem } from '@agentmem/sdk';

const client = new AgentMem('am_live_xxx...');

// Store
await client.memories.store('User likes pizza');

// Search
const results = await client.memories.search('food preferences');

// List
const all = await client.memories.list({ limit: 100 });
        

Python

          pip install agentmem

from agentmem import AgentMem

client = AgentMem('am_live_xxx...')

# Store
client.memories.store('User likes pizza')

# Search
results = client.memories.search('food preferences')

# List
all_memories = client.memories.list(limit=100)
        

SDKs coming soon. For now, use the REST API directly.