Getting Started

Quickstart

Two ways to give your agents memory: connect an MCP client with one config block, or install the Python SDK. Same API key, same memory — pick your path.

Get your API key

Create a free account and mint your API key instantly from the playground. Keys follow the format lk_... — the same key works for both paths.

Create an account

Add Litica to your MCP client

Open your Claude Desktop config file and add the following. Replace lk_your-api-key with the key you received.

~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "litica-memory": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.litica.org/mcp/",
        "--header",
        "X-API-Key:lk_your-api-key"
      ]
    }
  }
}

Save the file and restart Claude Desktop. That is it.

Clients that speak HTTP natively (Claude Code, Cursor) can skip the mcp-remote bridge and point straight at the server:

{
  "mcpServers": {
    "litica-memory": {
      "url": "https://mcp.litica.org/mcp/",
      "headers": { "X-API-Key": "lk_your-api-key" }
    }
  }
}

You are connected.

Claude now has Litica's four memory tools: add_memory, search_memory, add_memories_batch, and add_documents. Tell Claude to remember something and it saves it, broken down by what it's about. Start a new session, ask it to recall, and it brings the right memories back, most useful first.

Next steps

Configuration guide for Claude Code, Cursor, and other MCP clients
MCP Server reference for tool parameters and how retrieval works

Install the package

Requires Python 3.10+. The only dependency is httpx.

pip install litica

Get your API key

Create a free account and mint your API key instantly from the playground. Keys follow the format lk_... — the same key works for both paths.

Create an account

Write a memory, then search it

Writes are queued: the server accepts them instantly, then decomposes, embeds, and stores them in the background. So a first script polls briefly between write and search:

first_memory.py

import time
import litica

client = litica.Client(api_key="lk_your-api-key")

client.add_memory("Sam owns the Atlas pricing page.")

# Poll until the write has landed (usually a few seconds).
for _ in range(20):
    hits = client.search_memories("who owns pricing?")
    if hits:
        break
    time.sleep(3)

for hit in hits:
    print(hit.id, hit.text)

That is the whole loop.

Your agent now writes memories that persist across sessions and retrieves them ranked by what fits the task. Namespaces, tracing, and the audit feed all build on these two calls.

Next steps

Full SDK quickstart — agent identity, environment variables, and the read-your-writes pattern
Configuration — scope defaults, timeouts, and client lifecycle
SDK API Reference — every method, parameter, and response type