Choose a prompt
Website prompt
New to Synap or not signed in? Use the public prompt to let your coding agent assess your project before setup begins.
Dashboard prompt
Already signed in to Synap? Use the dashboard prompt to begin setup with your existing project context.
Website prompt
Use this public prompt if you are exploring Synap, have not created a project yet, or are not signed in. It gives your coding agent the context to assess your app and help you decide whether to set up Synap. Paste this prompt into your coding agent:This is the same prompt copied from www.maximem.ai. It is a public starting point, so it cannot access your Synap account or project.
Dashboard prompt
Use the dashboard prompt if you are signed in to Synap. It is prepared for your existing setup, so your coding agent can create or reuse the right instance and configure your app without you handling API keys.Open Dashboard setup
Generate and copy a prompt configured for your Synap account and project.
- In the Dashboard, open Setup → Coding agent and copy the generated prompt.
- Paste it into the coding agent working in your project.
- Let the agent review the project and ask the decisions it needs before it changes anything.
What each prompt does
The website prompt helps your agent understand Synap, inspect your project, and explain what Synap would change. It asks for your approval before it starts setup. The dashboard prompt begins the authenticated setup flow. The agent signs in with the included one-time code, checks for existing instances, asks whether to reuse or create one, creates an app API key, and writes the integration. It waits for the memory architecture to be ready before it tests reads or writes.Dashboard prompt access and expiry
The dashboard setup code is single-use and expires after 10 minutes. Generate a fresh prompt if it expires or has already been used. A code works only in the region where it was created, so copy it from a browser on the same network as the agent’s machine. The setup key can read and create instances, read and create API keys, and read use cases. It cannot access billing or team members, delete resources, or perform other account administration. It lasts 7 days and cannot be renewed withauth refresh. It can revoke only itself and API keys it minted. The agent should run auth logout after a successful setup; that revokes the setup key while leaving the runtime key in your app’s .env file.
This differs from a normal browser login: browser-login keys last 90 days and can be renewed with maximem-synap auth refresh before expiry.
What the agent will ask and do
The agent checks whether an instance already exists for this application. It asks whether to reuse it or create a new one; an instance normally serves one agent and one memory design. It also asks for the choices needed to generate the use-case file and memory architecture. It does not commit changes. It asks before creating an instance or making application changes. Expect the memory architecture build to take about 15 to 20 minutes. Until it is ready, Synap rejects runtime reads and writes with HTTP 409instance_not_ready.
Continue the conversation
Once setup is complete, ask your coding agent what you want to build:Add long-term memory to my LangGraph agent, scoped per user.
Store customer preferences and retrieve them in my support agent.
Show me where Synap is initialized and explain the retrieval flow.
Install the skill yourself
The website prompt installs the public skill for you. If you want to add it yourself for durable Synap instructions across future projects and conversations, use the command below. The skill teaches your agent the SDK, scoping model, and framework integrations.Claude Code
~/.claude/skills/synap instead. For other tools, copy skills/synap/AGENTS.md to the tool’s rules-file location; see the skill README.
Next steps
Quickstart
Learn the runtime concepts behind the integration.
Instances
Decide when to create, reuse, or promote an instance.