# Getting Started with TeamDay Set up your TeamDay account, create your first AI agent, and start chatting in minutes. # Getting Started with TeamDay This guide walks you through signing up, creating your first AI agent, and having your first conversation. You'll be up and running in a few minutes. --- ## Sign Up 1. Visit [teamday.ai](https://teamday.ai) and click **Sign Up** 2. Authenticate with **Google**, **GitHub**, or **Email** 3. An organization is created for you automatically Once signed in, you land on the org home page where you can create agents, open Spaces, and start chatting. --- ## Core Concepts Before building, here's how TeamDay is organized: | Concept | What It Is | |---------|-----------| | **Agent** | An AI employee — a specialized team member with a name, role, system prompt, model, and equipment (skills, MCPs, subagents). | | **Space** | A workspace (project directory) where agents run. Contains files, installed tools, skills, and secrets. | | **Skill** | A reusable capability that teaches agents how to do specific tasks (e.g., generate images, write reports). | | **MCP Server** | A tool integration that connects agents to external services (Google Analytics, Ahrefs, Slack, etc.). | | **Mission** | A scheduled task — run an agent on a cron schedule, once at a future time, or continuously. | --- ## Onboarding Flow The recommended path to get started: 1. **Sign up** and create your organization 2. **Create an agent** — your first AI team member 3. **Chat at org level** — test the agent from your org home page 4. **Create a Space** — set up a workspace for a project 5. **Add the agent to the Space** — give it access to the Space's files and resources --- ## Create Your First Agent ### In the Web App 1. From the org home page, click **+ Create Agent** 2. Set the basics: - **Name**: "My Assistant" - **Role**: "General Assistant" - **System Prompt**: "You are a helpful assistant. Be concise and friendly." 3. Start chatting immediately ### Via CLI ```bash teamday agents create \ --name "My Assistant" \ --role "General Assistant" \ --system-prompt "You are a helpful assistant. Be concise and friendly." ``` ### Via API ```bash curl -X POST "https://cc.teamday.ai/api/v1/agents" \ -H "Authorization: Bearer $TEAMDAY_API_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "name": "My Assistant", "role": "General Assistant", "systemPrompt": "You are a helpful assistant. Be concise and friendly." }' ``` --- ## Chat With Your Agent Agents can be chatted with at two levels: ### Organization-Level Chat Chat with an agent directly from the org home page. The agent uses only its own equipment — skills, MCPs, and subagents attached to it. No Space context is involved. This is great for quick conversations, general questions, and tasks that don't require access to project files. ### Space-Level Chat Chat with an agent inside a Space. The agent uses its own equipment **plus** the Space's installed resources (skills, MCPs, secrets, files). This is the union of both — everything the agent has, plus everything the Space provides. Use Space-level chat when the agent needs to work with files, run commands, or access Space-specific integrations. ### Built-in Tools Agents in TeamDay are powered by Claude and have access to a full set of tools: - **Read, Write, Edit** — work with files in the Space - **Bash** — run shell commands - **WebSearch, WebFetch** — search the web and fetch pages - **Glob, Grep** — find and search files These tools let agents do real work — not just answer questions. ### Via CLI Interactive multi-turn chat: ```bash teamday agents chat ``` Single message execution: ```bash teamday agents exec "Summarize the README in this project" ``` Add a Space context so the agent can access files: ```bash teamday agents exec "Review the code in src/" --space ``` --- ## Create a Space Spaces are project workspaces where agents do their work with files, commands, and integrations. 1. Click **+ New Space** from the dashboard 2. Give it a name (e.g., "My First Project") 3. Choose a type: - **Empty** — blank workspace - **Git** — clone from a GitHub repository 4. Click **Create** Your Space starts with a sandbox directory where agents can read and write files, run commands, and use installed tools. **Via CLI:** ```bash teamday spaces create --name "My First Project" ``` --- ## Equip Your Agent Agents have four types of equipment, managed via the agent's detail panel: | Equipment | What It Does | |-----------|-------------| | **Tools** | Built-in Claude tools (Read, Write, Edit, Bash, Glob, Grep, WebSearch, WebFetch) — available by default | | **Skills** | Reusable capabilities (e.g., research, data analysis, image generation) | | **MCPs** | External integrations (Google Analytics, Ahrefs, databases, etc.) | | **Subagents** | Other agents this agent can delegate work to | ### Attach Skills ```bash teamday agents add-skill core:research-assistant core:data-analyst ``` Or attach them through the agent's detail panel in the web app. ### Connect MCP Servers MCP (Model Context Protocol) servers connect agents to external tools and services: ```bash teamday agents add-mcp google-search ``` Popular integrations include Google Analytics, Ahrefs, Slack, and GitHub. See [MCP Servers](https://docs.teamday.ai/guides/mcp-servers/what-are-mcp-servers) for the full list. ### Store Secrets If your MCPs or skills need API keys, store them as encrypted Space secrets: ```bash teamday spaces set-secret OPENAI_API_KEY=sk-abc123 ``` --- ## Visibility vs Attachment These are two separate concepts: - **Visibility** controls who can **see and discover** an item: `private`, `organization`, `public`, or `unlisted`. - **Attachment** controls where an item is **active**: attached to a specific agent or Space. Setting an agent's visibility to `organization` means all org members can see it — but it does **not** mean the agent is automatically included in every Space or chat. Items must be explicitly attached to an agent or Space to be used. --- ## Set Up Authentication For API and CLI access, you need credentials. ### Option 1: OAuth Login (CLI) ```bash teamday auth login ``` Opens your browser for secure authentication. Tokens are managed automatically. ### Option 2: Personal Access Token (API / CI) 1. Go to **Settings** in the web app 2. Create a new **Personal Access Token** 3. Copy and save it immediately (shown only once) ```bash # Use with CLI teamday auth set-key "td_your_token_here" # Or as environment variable export TEAMDAY_API_TOKEN="td_your_token_here" ``` See [API Keys & Authentication](https://docs.teamday.ai/guides/api-keys) for the full guide. ### Claude API Credentials TeamDay agents are powered by Claude. By default, the platform provides Claude API access. You can also bring your own credentials for direct billing: ```bash # Check what's configured teamday keys status # Set your Claude OAuth token (from claude.ai subscription) teamday keys set oauth sk-ant-oat01-your-token # Or set an Anthropic API key teamday keys set api-key sk-ant-api03-your-key ``` TeamDay uses a priority system: your personal credentials are used first, then organization credentials, then the platform default. --- ## Next Steps | Guide | What You'll Learn | |-------|------------------| | [Create Your First Agent](https://docs.teamday.ai/guides/first-agent) | In-depth agent creation with equipment, models, and system prompts | | [Spaces & Workspaces](https://docs.teamday.ai/guides/spaces) | Configure project directories, files, and Git repos | | [Skills](https://docs.teamday.ai/guides/skills) | Build and install reusable agent capabilities | | [MCP Servers](https://docs.teamday.ai/guides/mcp-servers/what-are-mcp-servers) | Connect to external tools and services | | [Missions](https://docs.teamday.ai/guides/missions-user-guide) | Schedule agents to run tasks automatically | | [CLI Tool](https://docs.teamday.ai/guides/cli-tool) | Full command-line reference | --- ## Get Help - **Support:** support at teamday.ai - **Community:** [Discord](https://discord.com/invite/DbUyE6b4t5) - **API Reference:** [/docs/api](https://docs.teamday.ai/api)