Mission Squad Platform UI

This section documents the Mission Squad Platform UI. It covers dashboard navigation, managing providers/models/agents, MCP tools, embeddings (vector stores and files), workflows, schedules, security, and troubleshooting. Each page includes API parity links to keep UI and API usage aligned.

• Getting Started
• API docs
• Endpoint Index

Guides

• Providers — connect and manage upstream AI providers
• Models — discover vendor models and save named configurations
• Agents — create agents (prompts, models, tools, RAG)
• MCP (Tools) — connect tool servers and inspect available functions
• Prompt Studio — iterate prompts, stream events, save back to agents
• Embeddings — upload files, manage vector stores, diagnostics
• Workflows — compose multi‑agent pipelines
• Schedules — run agents/workflows on a cadence with delivery
• Recipes — opinionated end‑to‑end examples
• Security & Isolation — guarantees and best practices
• Troubleshooting — common issues and quick checks
• Hosting — deploy the Platform API and MCP via Docker & Compose

Concepts at a glance

• Providers: secure, per‑account credentials for model vendors (OpenAI, Google, Anthropic, Groq, OpenAI‑compatible runtimes).
• Models: named configurations per provider model (temperature, max tokens, etc.).
• Agents: model + system prompt with optional Tools (MCP) and RAG collections.
• MCP: Model Context Protocol servers that expose callable tools.
• Prompt Studio: scratchpad to test agents/models and watch tool events.
• Embeddings: files and vector stores used for RAG.
• Workflows: multi‑agent orchestration (helpers → primary).
• Schedules: recurring agent/workflow runs with delivery.