# Sculptured > Sculptured lets you design a team of agents performing complex tasks on a visual board, run it and see the evidence behind every result. ## Get started - [Introduction](/introduction.md): Design, run and supervise teams of agents performing complex tasks. - [Quickstart](/quickstart.md): Set up Sculptured, create a team and follow its first run. - [Set up your account](/guides/setup.md): The setup steps after sign-up on the hosted service, how model keys are kept, and what key-based runs cost. - [Open models with OpenCode](/guides/open-models.md): Run agents on open models such as Kimi, GLM, DeepSeek or GPT OSS, through OpenCode and an OpenRouter key. - [Connect a computer](/connect-a-computer.md): Install Sculptured's runner with one command, choose which folders agents may use, and keep it running in the background. - [Concepts](/concepts.md): The words Sculptured uses, how the parts connect, and what each status means. ## Design a team - [Designing teams of agents](/guides/agent-design.md): Choose roles, define handoffs and give each agent a clear job, a budget and a way to check its work. - [Compose on the board](/guides/board.md): Choose a folder, add agents and connect handoffs on the Design board. - [Design assistant](/guides/design-assistant.md): Ask your own Claude Code or Codex to change a team's design, inspect the proposal on the board, then approve or reject it. - [Examples and templates](/guides/examples.md): Every starting point that ships with Sculptured — what each agent does, how work flows, and when to pick it. - [Reusable groups](/guides/groups.md): Save a group of agents once in your Library and add an independent copy to any project — with its jobs, arrows and workflow stages. ## Run a team - [Run a team](/guides/running.md): Start a team, follow its work on the Run board, and pass work along again when you need to. - [Control live agents](/guides/control.md): Message an agent, interrupt its current turn, request a summary or stop its session. - [Budgets and limits](/guides/budgets.md): Keep a run within a share of your Claude or Codex plan, a token count or a time, with a soft value that wraps up and a hard value that stops. - [Where tokens went](/guides/where-tokens-went.md): See what a run's tokens were spent on (setup, files, handoffs, the plan window) and what to change to use less. - [Send traces](/guides/send-traces.md): Send a team's runs as OpenTelemetry traces to Laminar or another collector you already use. - [Run from a terminal](/guides/run-from-terminal.md): Run a team from its file with one command, see what happens as it happens, and use the exit code in scripts. - [Questions and approvals](/guides/questions.md): What happens when an agent asks you something, how long it waits, and what it does if nobody answers. - [Context and communication](/guides/context.md): Inspect captured context, tool results, messages and usage, with the source of each claim. - [The status page](/guides/status-page.md): One page an agent keeps current with where the work stands and what changed, under Results → Status. - [Changes to business systems](/guides/changes.md): Agents propose each change to an order, a schedule or another system first. Sculptured checks your rules itself, then approves the change or asks you. - [Events from other systems](/guides/triggers.md): Give a team an address that an ERP, a maintenance system or a line's alarms can call. Each event reaches an agent, and the team starts if it isn't running. - [Result checks](/guides/result-checks.md): Rules an agent's result must pass before its work is handed on. Sculptured checks them itself and sends a failing result back. - [What agents know](/guides/knowledge.md): The project map made at every start, lessons agents record, a search of the team's history, and the Knowledge view. - [Skills and tools on the agent's computer](/guides/skills.md): Give an agent skills it loads only when it needs them, and tools installed on its computer, such as Kindex. - [Git workflows](/guides/git.md): Review what agents changed, commit it deliberately, and merge each agent's own copy when they work in parallel. - [Connection interruptions](/guides/offline.md): What keeps going, what waits and what happens on reconnect when the computer running your agents loses its connection mid-run. ## Build workflows - [Workflows](/guides/workflows.md): Route work items through stages with explicit outcomes, review decisions, checks and revision limits. - [Textbook creator](/guides/textbook.md): A worked dispatch workflow — turn sources into interactive textbook sections, each researched, built, checked twice and assembled into one book. - [CS2680 pattern catalog](/guides/pattern-catalog.md): Eight small Community blocks, one per multi-agent architecture from the CS2680 Agent Design II lecture. ## Save and share - [Team files](/guides/team-files.md): Save a team or a group as a file, open it in another Sculptured, and edit it with completion in your editor. - [MADL files](/guides/madl.md): Export a team as a Multi-Agent Descriptor Language file that other tools can read, and import MADL files as teams. - [Import a Claude Code or Codex session](/guides/import-a-session.md): Turn a session you ran outside Sculptured into a run you can replay in History and a team you can run again, subagents included. - [Community](/guides/community.md): Find ready-made groups of agents and the people who made them, save the versions you use, and publish your own. ## Reference - [CLI reference](/reference/cli.md): Every command and flag of the sculptured program, with defaults and examples. - [Agent tools](/reference/agent-tools.md): The tools Sculptured gives every Claude Code and Codex agent, what each one does, and what you see when an agent uses it. - [Older team files](/reference/team-file.md): sculptured.team/v1, the JSON format Sculptured saved teams in before team files became MADL documents. - [States and evidence](/reference/states.md): Every state an agent, a question and a result can be in, what each one means, and what you can do next. - [Limits and usage](/reference/limits.md): Token limits, agent counts, usage holds and timeouts, with their defaults and enforcement behavior. ## Hosting and security - [Run it yourself](/operations.md): Build Sculptured from source, run it on your own computer, and back up, upgrade and roll back your data. - [Security](/security.md): Where your code, credentials and agents live, what the server can ask your computer to do, and what your computer refuses. - [Privacy](/privacy.md): Understand the difference between usage analytics, stored run history and data on your computer. ## Help - [Troubleshooting](/troubleshooting.md): Fixes for the problems people run into most, grouped by where they show up. - [FAQ](/faq.md): Short answers about what Sculptured is, what it costs to run, where it runs and what happens to your data. - [Changelog](/changelog.md): What changed in Sculptured, newest first.