Alfe developer docsGive your agentsomewhere to live.

Choose your path

Start with the route that matches how you want to bring an agent onto Alfe.

Run it yourselfInstall the CLIConnect an agent from your own machine or server with @alfe.ai/cli.Open the quickstart →Agent-led setupBootstrap over MCPLet an MCP-capable agent discover Alfe and onboard itself.Explore MCP →Build with codeUse the agent APIGive agents self-service access to memory, identity, sync, and integrations.View the API →Reach the worldAdd channels and toolsConnect each agent to the services and conversations it needs.Browse integrations →

Understand the platform

What is Alfe?The agent operating system and the problems it solves.

Core conceptsRuntime, memory, identity, model routing, and credits.

Managed agentsA ready-to-run home for an agent, operated by Alfe.

Agent runtimesChoose the runtime your agent runs on, and switch later.

Team packsLaunch a whole team of agents from one template pack.

OrganizationsMembers, roles, teams, projects, and shared knowledge.

Give your agent more to work with

Scheduled jobsRun a prompt on a schedule and deliver the result to a conversation.

Remote browser & terminalWatch your agent's browser, take over, and open a terminal.

GPU computeLet an agent rent GPUs on your own Vast.ai account.

Persistent memoryWhat your agent remembers across sessions, and how to manage it.

ChannelsReach your agent from Slack, Teams, WhatsApp, phone, and more.

WebhooksWake your agent when something happens in another system.