Concepts

The vocabulary every other page links back to, and a six-step orientation to how Trinity works end to end.

Glossary

Autonomous AgentAn AI system that plans and executes tasks independently. Each agent runs as an isolated Docker container with pre-installed languages (Python 3.13, Node.js 20, Go 1.23) and a pluggable agent runtime: Claude Code, OpenAI Codex, or Gemini CLI (see Agent Runtimes). Agents persist memory across sessions, delegate to other agents, and run on schedules without human intervention.

Agent ContainerAn isolated Docker container with standardized interfaces for credentials, tools, and MCP server integrations.

TemplateA GitHub repository or local directory that defines an agent's initial configuration, including CLAUDE.md, template.yaml, .mcp.json.template, and credential declarations.

MCPThe Model Context Protocol — the standard agents use to communicate with each other and with external tools. Trinity's MCP server exposes 129 tools for fleet management, credential injection, scheduling, file sharing, per-user memory, channel messaging, and more. See MCP Integration.

System AgentAn auto-deployed platform orchestrator (trinity-system) that manages fleet operations such as health checks, scaling, and coordination.

Autonomy ModeA master toggle that enables or disables all scheduled operations for a given agent.

ExecutionA single run of a task on an agent. Executions can be triggered manually, by a cron schedule, by another agent, or via the API. See Executions.

Public LinkA shareable URL that allows unauthenticated users to chat with an agent directly. See Public Links.

WorkspaceThe chat app. Multiple conversations per agent that keep memory across turns, with voice, file drops, loops, deliverables, and the agent's canvas in one place. Platform users open it from the top nav; external clients sign in to it directly. See Workspace.

CanvasA surface an agent keeps current — blocks of text, metrics, charts, images, and diagrams it updates as it works — shown beside the conversation. See Agent Canvas.

How It Works

Trinity runs as a set of Docker containers on your local machine or server. After starting the platform you interact with it through the web UI or the API.

1
Start the platform
Run ./scripts/deploy/start.sh to bring up all services. The web UI is available at http://localhost and the API at http://localhost:8000/docs. On your first login a short first-run setup connects Claude and walks you to your first agent. See First-Time Setup.
2
Create an agent
From the dashboard, click Create Agent and select a template (GitHub repo URL or local path). Trinity pulls the template, builds a container, and deploys the agent. See Building Agents.
3
Configure credentials
Add API keys and secrets through the agent's credential panel. Credentials are encrypted and injected into the container at runtime with hot-reload support. See Credential Management.
4
Chat with the agent
Open the Workspace for a continuous conversation, or use the stateless Chat tab on the agent detail page for one-off turns. The agent processes your request using its configured tools, MCP connections, and reasoning context.
5
Schedule autonomous work
Set up cron-based schedules so the agent executes tasks on its own. Enable Autonomy Mode to let the agent's enabled schedules run. See Scheduling.
6
Monitor the fleet
Use the Dashboard (Timeline, Grid, or List) to view agent health and execution history; inter-agent calls show up in the Timeline replay. See Dashboard and Monitoring.

Going Deeper

System Overview — component table, ports, and the technology stack.

Authentication — JWT tokens, API keys, and the full REST surface (500+ endpoints).

Multi-Agent Collaboration — how agents delegate to each other, subscribe to events, and share rooms.

How Trinity Works — a 4½-minute animated film of one scheduled task end to end.

Building Agents