Documentation

dispatchmy.ai docs

Run the local agent runtime, connect it to the dashboard, and build workflows where a manager agent delegates to specialists with their own prompts, models, tools, and memory.

Overview

dispatchmy.ai pairs a browser dashboard with a local daemon. The daemon stores agent configuration, runs each agent in its own Docker container, and proxies upstream credentials so most secrets never reach the agents themselves.

Dashboard

Use the dashboard to create agents, connect tools, inspect runs, and manage the daemon.

Local runtime

The daemon runs on your machine, launches an isolated container per agent, and proxies provider keys server-side.

Delegated workflows

A manager agent calls specialists as tools — each one focused on a single job, with its own prompt, model, and tool set.

Tooling layer

Plug in model providers, MCP servers, and built-in tools so agents can act on real systems instead of just producing text.

Getting started

Start here if you're setting up dispatchmy.ai for the first time. These guides walk you from sign-in to a paired daemon and your first working agent.

Quickstart

Sign in, start the daemon, pair it with the dashboard, and run a first agent.

Installation

Supported platforms, Docker prerequisites, and what the dashboard's one-shot install command does.

Pairing the daemon

How daemon pairing works at a high level — pair codes themselves only appear in the signed-in dashboard.

Agents

Agents are the building blocks of a workflow. Each one has a prompt, model, tool set, optional memory, and optional schedule.

Agent model

Agents, subagents, and sessions — and how delegation works.

Creating agents

Start small, name agents by responsibility, and add complexity once the first run works.

Memory

What memory is, when to enable it, and when an MCP server is the better fit.

Schedules

Run an agent on a recurring cadence and review what each run did.

Runtime

The daemon runs agents on your machine; the dashboard is the browser interface that controls it. These guides cover how they fit together, how updates work, and where data lives.

Local daemon

What the daemon is responsible for, why it needs the Docker socket, and what stays local.

Dashboard

How the dashboard controls agents, picks the daemon to talk to, and why it requires sign-in.

Updates

API version checks, the one-click upgrade flow, and recovery if an update fails.

Data storage

Where configs, logs, sessions, memory, and artifacts live on the host.

Tools and integrations

Connect model providers, MCP servers, and built-in tools so agents can browse, edit files, call APIs, and work with your existing services.

Model providers

Anthropic, OpenRouter, and custom OpenAI-compatible endpoints — and how keys are stored.

MCP servers

The curated catalog vs. custom entries, supported transports, and tool filters.

Built-in tools

First-party packages — browser, shell, file editing, GitHub, web fetch and search — and their container dependencies.

Secrets

What the daemon proxies, what reaches the agent container, and how to rotate a key.

Operations

Keep your setup healthy after the first successful run. Diagnose connection issues, inspect logs, understand the security model, and know where to ask for help.

Troubleshooting

Daemon offline, model-provider errors, and Docker-side failures.

Logs

Where logs appear, what they contain, and what to redact before sharing one.

Security model

Local execution, Docker isolation, credential proxying, and dashboard authentication.

Support

What to include in a bug report, what to keep out, and how to reach us.

Reference

Look up exact behavior when a walkthrough isn't what you need — what an agent's config holds, where the daemon API stands, what platforms are supported.

Configuration reference

What an agent's configuration holds and where it's edited.

Limits

Supported platforms, the Docker requirement, and known model-provider caveats.

Changelog

Where release notes will live once the beta cadence stabilizes.