⚒ Use cases

What squads get deployed to do.

Real patterns from the platform: compose agents into workflows and factories, connect the tools you already use, and read the trace of every run.

  • A YouTube topic-research factory

    Research pipelines

    Describe the pipeline in plain English — "search YouTube for a topic, pull transcripts for the top five videos, write one holistic summary" — and the configuration agent proposes the agents, models, and hand-offs, then builds the factory. Each step's output feeds the next, and any step can be a whole workflow.

  • A nine-agent financial analysis

    Multi-analyst reports

    One request fans out to specialist agents — fundamentals, sentiment, macro, and six technical analysts — each pulling its own data through MCP tools and writing its own section, before a lead agent stitches them into one report. Every tool call, argument, and returned payload stays in the trace.

  • Your Tuesday, handled before you sit down

    Scheduled operations

    Put workflows on a schedule so recurring work — status rollups, inbox triage, data pulls, weekly summaries — runs with a full trace while you do the work only you can do. When a run does something unexpected, you read exactly what happened instead of guessing.

  • RAG over your documents

    Assistants grounded in your data

    Create vector stores, upload your files, and give agents retrieval over your private data. Squads answer from your documents and act in your systems — Google Workspace, Notion, Asana, Shopify, and dozens more connectors, or any MCP server you bring.

  • Your VPC, your metal, your keys

    Private deployments

    Run the whole platform in your own VPC or on-prem, pointed at any model provider — or open models on hardware you own. Swap the model behind any agent in seconds with no rewrite, and keep keys, data, and traffic under your control.