> ## Documentation Index > Fetch the complete documentation index at: https://docs.maximem.ai/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > Maximem Synap is a managed memory layer for AI agents: it ingests conversations and documents, extracts structured memory, and returns scoped context to the agent at retrieval time. > The Python and JavaScript/TypeScript SDKs are the supported way to use Synap. Write integration code with SDK calls, not raw HTTP requests. > Start with the Quickstart: https://docs.maximem.ai/getting-started/quickstart.md. Every SDK method is documented in the SDK reference: https://docs.maximem.ai/sdk-reference/overview.md. # Vity or Synap? > Which Maximem product fits what you are doing, which dashboard to sign in to, and why their keys and memory do not mix. Maximem makes two products, and they do not share accounts, keys, memory or credits. One question decides which you want: are you using AI tools, or building one for other people? * **Vity** is memory for **you**, in the AI tools you use: ChatGPT, Claude, Gemini and Perplexity in your browser, OpenClaw, Hermes, and MCP clients such as Claude Code and Cursor. * **Synap** is memory for **your users**, inside an agent or app you build, through an SDK, a framework integration, or the Synap MCP server. ## Side by side | | Vity | Synap | | - | - | - | | **Who it's for** | People using AI apps and agents | Developers building an agent or app for others | | **Whose memory** | Yours: one memory per Vity account | Each of your app's users, kept apart | | **What you install** | A browser extension, the OpenClaw or Hermes plugin, or an MCP server entry in your client | The Python or JavaScript SDK, a framework integration, or the Synap MCP server | | **Dashboard** | [app.maximem.ai](https://app.maximem.ai), sign in with Google | [synap.maximem.ai](https://synap.maximem.ai), with its own sign-in | | **API key** | Starts with `mx_`. You create it under **API Keys** in the Vity dashboard. The browser extension needs no key. | Starts with `synap_`. You create it for an agent instance in the Synap dashboard. | | **Credits** | A Vity credit balance, shown on the Vity dashboard | Synap plans and credits, managed in the Synap dashboard | ## Keys only work in their own product Each product checks a key against its own list of keys, so a key from one is unknown to the other. * **A Synap key in a Vity tool** is rejected with a message that says it is a Synap key and points you to [app.maximem.ai/api-keys](https://app.maximem.ai/api-keys) for a Vity one. * **A Vity key in a Synap app** is rejected as invalid credentials. Synap does not say which product the key came from, so check the prefix yourself: `mx_` belongs to Vity. If a tool says your key is invalid, look at the first few characters before anything else: a key from the wrong dashboard will never work, however carefully you copy it. ## Memory is not shared Vity stores and recalls your memory through its own services, not through Synap. A Synap app reads only memory stored in Synap. So what Vity has learned about you is not visible to an agent you build on Synap, and the reverse. Inside Vity, everything you connect shares one memory: every API key on your account and your browser extension sign-in reach the same place. To keep two sets of memories apart in Vity, use two Vity accounts; a second key on the same account reaches the same memory. ## Accounts and credits are separate Each product has its own sign-in. A Vity account does not give you a Synap account, or the other way round. Credits are separate too. Vity keeps its own balance, and your first Vity API key adds 10,000 credits to it. Synap bills through its own plans and credits; see [Pricing](/resources/pricing). Credits in one product cannot pay for the other. ## Which one fits you ### I use ChatGPT and Claude in my browser Use Vity. Install the [browser extension](/vity/browser-extension) and sign in with Google; no key needed. Your chats on one AI site become memory the extension can draw on when you use the others. ### I use Claude Code or Cursor Use Vity's [MCP server](/vity/mcp/overview) with an `mx_` key if you want the agent to remember **you**: your preferences, your projects and how you write, across sessions and across tools. If you are using Claude Code to **build** an agent, and that agent needs memory for its own users, that agent uses Synap: start with the [Synap quickstart](/getting-started/quickstart). Many developers use both: Vity for themselves, Synap inside what they ship. > [!NOTE] > > The Vity MCP server and the [Synap MCP server](/integrations/mcp) are different servers. The Vity one takes an `mx_` key and serves your own memory. The Synap one takes a `synap_` key and serves the users of an agent you build. ### I'm building a support bot Use Synap. A support bot talks to many customers, and each one needs their own memory. Synap keeps each user's memory separate, and on a business (B2B) instance it also groups users under the customer they belong to. Vity is the wrong fit here: a Vity key belongs to one person's account, so everything the bot saved would land in a single memory, mixed across all your customers. Start with the [Synap quickstart](/getting-started/quickstart), or see [integrations](/integrations/overview) for your framework. ### I run OpenClaw or Hermes Use Vity. Install the [OpenClaw plugin](/vity/openclaw) or the [Hermes plugin](/vity/hermes) with an `mx_` key, and your agent remembers you across sessions. The [Quickstart](/vity/quickstart) has the setup command for each. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.