This guide walks you through your first five minutes with simse: connecting a backend, starting a session, and seeing persistent context in action.
Prerequisites: simse installed and available in your PATH. An API key for at least one AI backend (OpenAI, Anthropic, or any ACP-compatible service).
simsesimse opens a terminal UI. All commands are entered as slash commands by typing / in the input prompt.
simse uses ACP (Agent Client Protocol) to connect to AI backends. Add your first backend to ~/.config/simse/acp.json:
{
"servers": [
{
"name": "anthropic",
"command": "npx",
"args": ["-y", "@anthropic-ai/claude-code"],
"env": { "ANTHROPIC_API_KEY": "your-anthropic-api-key" },
"defaultAgent": "claude-sonnet-4-6"
}
],
"defaultServer": "anthropic"
}Restart simse to pick up the change. Use /status in the TUI to verify the server is running.
You are already in a session when simse launches. Ask simse something about a project you are working on. The more context you provide upfront, the less you will need to repeat later.
End the session with Ctrl+D or by typing /exit. Then start a new session by launching simse again:
simseAsk simse about the same project. The context from your first session carries over automatically. simse surfaces relevant entries from the adaptive store based on the current session.
Extend simse by connecting an MCP server. Add it to ~/.config/simse/mcp.json:
{
"servers": [
{
"name": "filesystem",
"transport": "stdio",
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/project"]
}
]
}Restart simse to load the new tool. Tools appear in the tool registry and are available in all sessions.
The library accumulates knowledge from your sessions automatically. Ask simse to search it:
Search the library for notes about my project.
The library stores, catalogs, and deduplicates context automatically. Over time, it becomes a searchable record of your working knowledge.
/fork to explore a different approach while keeping shared context