Skip to content

quickstart.mdx

Latest commit

 

History

History
80 lines (55 loc) · 3.84 KB

quickstart.mdx

File metadata and controls

80 lines (55 loc) · 3.84 KB
Raw
title Quick Start
description Experience AI memory in five minutes — chat, upload, retrieve, and connect your tools

This guide takes you from sign-up to a working memory in about five minutes, then points you at the integration path that fits how you work.

Step 1: Sign Up

Create an account at app.memorylake.ai. MemoryLake automatically provisions your personal workspace and assistant — there is nothing to configure before your first conversation.

Step 2: Experience Memory in the Playground

Open the Playground from the sidebar (new accounts land there with a short guided tour):

Try: *"Remember that my name is Ming and I prefer answers in bullet points."*

Behind the scenes, MemoryLake extracts a durable memory from the conversation — not a transcript, a fact.

Subscribe your project to a curated industry dataset — research papers, financial data, clinical trials, and more — so the assistant can search authoritative external sources alongside your own content. Attach a PDF, Word, Excel, or Markdown file directly in the chat. Its content is parsed, indexed, and becomes retrievable context. Start a **new** conversation and ask: *"What's my name and how do I like my answers?"* The assistant recalls it — that's memory persisting across sessions. The Home page's Memory On/Off comparison shows the same effect side by side. Visit **Memories** in the sidebar to see what was extracted — each memory shows its content and a **source trace** back to the conversation or document it came from.

Step 3: Create an API Key

For anything programmatic — the REST API, Model Router, or agent plugins — create an API key:

  1. Open API Keys in the console sidebar
  2. Click create, give the key a name, and pick an expiry (defaults to 30 days; "Never" is available)
  3. Copy the sk-… key when it is revealed — it is shown only once

See Authentication for header formats and key management.

Step 4: Put Memory Where You Work

Pick the path that matches your workflow:

Install the MemoryLake plugin — your agent automatically recalls context, captures new memories, and uploads files you share with it. Full programmatic control: projects, documents, memory add/search/trace, and conflict resolution at `https://app.memorylake.ai/openapi/memorylake`. Add memory to your existing OpenAI/Anthropic calls with a base-URL change. Private preview. Call mainstream LLMs through one OpenAI-compatible endpoint with the same API key.

What Just Happened?

Your conversation and file were distilled into memories inside a project in your workspace — retrievable by any assistant or API client you connect, forever scoped and private to you. When you're ready to share memory with a team, see Team Collaboration; to understand the full model, read Core Concepts.

Next Steps

  • Core Concepts — workspaces, projects, memories, actors, agents
  • Scenarios — personal memory, team knowledge bases, coding agents, and product integration walkthroughs
  • MemoryLake feature guide — documents, memories, conflicts, connectors, and more