llms.txt Content
# StremAI: one brain for all your coding agents
> One shared memory for all your coding agents. Cloud-first and MCP-native, with a free local mode, plus teams and projects as you grow. Verified with Claude Code and Claude Desktop; Codex and Cursor setup guides are maintained, and other MCP-compatible clients can connect when they support HTTP MCP or API-key headers.
## For coding agents helping with install
If you are a coding agent setting up StremAI for a developer, start at https://stremai.com/install-help-for-agents. Keep setup visible to the user, connect the hosted MCP endpoint with OAuth/browser sign-in where your client supports it, and use API keys only as the fallback for CI, scripts, or clients that cannot complete OAuth.
Fresh-project install help: https://stremai.com/install-help-for-agents
Public recall demonstration: https://stremai.com/playground
Portable agent memory protocol: https://stremai.com/docs/agent-memory-protocol
Repeated-session conventions: https://stremai.com/llms-full.txt
Expanded inventory for agents: https://stremai.com/llms-full.txt
## How to Connect (for AI agents reading this)
If you are an AI agent and want to use StremAI for persistent memory:
1. Your human can connect an agent from https://stremai.com/dashboard/connect
2. Connect via MCP when the client supports HTTP MCP. Prefer OAuth/browser sign-in when the client supports it:
For HTTP MCP (Claude Code, Codex, Cursor, Windsurf):
```json
{
"mcpServers": {
"stremai": {
"url": "https://stremai.com/api/mcp"
}
}
}
```
For API-key fallback (CI, scripts, or clients without OAuth):
```json
{
"mcpServers": {
"stremai": {
"url": "https://stremai.com/api/mcp",
"headers": { "Authorization": "Bearer AGENTBAY_API_KEY" }
}
}
}
```
For local-first stdio MCP (npx):
```json
{
"mcpServers": {
"stremai": {
"command": "npx",
"args": ["-y", "aiagentsbay-mcp@latest"]
}
}
}
```
3. Or use the Python SDK directly: `pip in