# Getting Started > Create AI-powered content with the Katchy API and MCP tools. --- # Analytics API > API reference for retrieving content analytics — dashboard metrics, post-level performance, and strategy insights. --- # Content > API reference for creating carousels, videos, social posts, and articles with Katchy. --- # Library > API reference for managing content in the Katchy library — list, retrieve, update, delete, and duplicate items. --- # Publish API > API reference for managing publications — list, update, delete, and trigger immediate publishing across social platforms. --- # Workspace API > API reference for workspaces, publication queue, and content inspirations. --- # Authentication > How to authenticate with the Katchy API using JWT or API keys. --- # Features > Overview of Katchy's AI-powered content creation, brand management, publishing, and analytics capabilities. --- # Create Your First Carousel > Step-by-step guide to creating an image carousel with the Katchy API and MCP tools. --- # Platform Integrations > Connect direct platform APIs for deep analytics, comment management, and audience insights. --- # Connect Publishing > Set up a publishing provider to post content to your social media accounts. --- # Premiers Pas > Creez du contenu propulse par l'IA avec l'API Katchy et les outils MCP. --- # MCP Setup > Connect Katchy to Claude Code via MCP for AI-powered content creation. --- # Claude Code Skill > Create a Katchy skill for Claude Code so it knows when and how to use Katchy tools. --- # MCP Tool Reference > Complete reference for all Katchy MCP tools — input schemas, descriptions, and usage notes. --- # Mental model in 5 minutes > Read this before clicking around. The product is dense; the model underneath is small. Five concepts and you can navigate everything. --- # Assets > API reference for guild assets — documents, images, videos, and other media. --- # Connectors & MCP > API reference for integration connectors and user MCP server endpoints. --- # Conversations & Messaging > API reference for creating conversations, sending messages, and managing real-time threaded discussions between humans and operators. --- # Guilds & Organizations > API reference for guild and organization endpoints — create, list, and manage virtual AI companies. --- # Operators > API reference for operator endpoints — create and manage AI team members with personas, traits, and capabilities. --- # Tasks & Plans > API reference for tasks, work plans, and operator requests. --- # Templates & API Keys > API reference for guild templates, API key management, and billing. --- # Work Items, Workspace & Notifications > API reference for managing work items (objectives, initiatives, epics, tasks), workspace documents, and user notifications. --- # Authentication > How to authenticate with the Guilde API using JWT or API keys. --- # Actions > The flat list of every tool method your operators can run — built-in, MCP, and integration-provided, side by side. --- # Connectors (moved) > Connectors were renamed. The same surface now lives as Actions (live verbs) + Integrations (installer). --- # Integrations > The installer. Browse the catalog, click Install, authenticate, and the integration's actions appear on the Actions page. --- # MCP servers > Use Model Context Protocol servers as a capability source. Bring your own tools, or pull from the open ecosystem. --- # Capability overview > Actions vs Integrations vs MCP vs Skills — what they are, when to use each, and why they're not the same thing. --- # Runtimes > Where operators actually execute — daemons, sandboxes, agent CLIs, terminals. The plumbing layer beneath your operators. --- # Skills > Named, installable bundles of capability. The unit you hire operators with. --- # Artifacts & canvases > The outputs operators produce — drafts, designs, plans, code. Versioned, indexed, and shareable. --- # Canvases > Visual artifacts your operators populate with live data — briefs, decks, dashboards. Mustache + frontmatter, refreshed by tool calls. --- # Conversations > Threaded chats with operators — when to use them, how they differ from work items, and what operators do during a thread. --- # Guilds > A guild is your AI company. Members + operators + workspaces under one roof. --- # Operators > AI workers with a real job — what they are, how they work, and how to get the most out of them. --- # Work items > Tracked deliverables — Objectives, Initiatives, and Tasks. The hierarchy your AI company actually plans against. --- # Workspaces > File storage roots operators read and write to. Yes — workspace means storage here, not your AI company. --- # Workstation > Long-lived cloud shells with agentproto and your connector keys pre-wired. Operators that need a real machine. --- # agentproto CLI > Install the agentproto CLI on your machine, sign in to Guilde, and run agent CLIs (Claude Code, Codex, OpenCode, Hermes, OpenClaw) under operator governance. --- # Billing & plans > Four plans, one currency — credits. BYOK works on every plan; bring your own LLM keys to bypass credit billing. --- # Browser > Three tabs — live sessions, the local-browser tunnel, and recordings. The browser-automation control panel for your guild. --- # Knowledge > Per-guild knowledge bases your operators can query. Hybrid graph + vector via GBrain, vector-only via Qdrant. --- # Notifications > Inbox for guild events — workflow completions, approval requests, mentions, operator alerts. 15-second polling, jump-to-source on click. --- # Secrets > Encrypted vault for credentials, API keys, and tokens that operators need but never see in plaintext. --- # Usage & BYOK > Credit balance, the live consumption ledger, and where you bring your own LLM keys to bypass credit billing. --- # Quickstart > From sign-up to your first operator reply in about a minute. --- # API keys > Programmatic access to Guilde — automate operator interactions, build integrations, run scripts. --- # Channels > Make operators reachable on Discord, Slack, and Telegram. DMs and @mentions route to the right operator. --- # Members > Invite humans to your guild. Roles control who can see what and who can change things. --- # Account > Your user-level profile and notification settings. Lives outside any guild — display name, email, avatar, orgs you belong to. --- # Brand > Design kit picker for the whole guild. Fork a built-in preset, edit its DESIGN.md, or pull a community kit from designkit.sh. --- # Guild settings > Three tabs — Tool Security (allowlists), Experimental (feature flags), Danger (archive, reset). Governance for the whole guild. --- # Organization settings > Org-level settings — five tabs covering identity, guilds, members, billing, and org-wide connector credentials. --- # Guild profile > Identity for your guild — name, slug, industry, timezone, description, and the org it belongs to. --- # Build a routine > Schedule a recurring operator job — daily summaries, weekly digests, on-demand reports. --- # Bring your own LLM keys (BYOK) > Add your Anthropic / OpenAI / Google / OpenRouter / Mistral / Groq key, route operator calls through it, bypass credit billing. --- # Connect a CRM > Wire HubSpot or Salesforce so operators can read deals, update contacts, and follow up on leads. --- # Connect Discord > Get your operators replying on Discord. DMs and @mentions in your server route to your guild. --- # Hire your first operator > Walk through hiring a single operator end-to-end and getting your first useful reply. --- # agentagencies/v1 > File-format spec for autonomous AI agencies — services, agreements, deliverables, invoices, routines. Companion site to AIP-8. --- # agentgovernance/v1 > Open file-format spec for approvals, audit logs, and autonomy policies. Vendor-neutral, third-party verifiable. Hash-chained. --- # What is Bureau > A real browser your agents can drive — stealth Firefox plus a local daemon that exposes browser tools as MCP over HTTP. --- # Account & AI credits > Register, buy credits, mint a bur_ key, and link the daemon for metered AI — or stay on your own Anthropic key. --- # Credentials > Store platform sign-in secrets in the OS Keychain, off the LLM path, so sessions can re-authenticate themselves. --- # Install > Install the Bureau CLI, meet the requirements, and get Camofox running. --- # MCP tools > The Bureau tool catalogue served over MCP, grouped by family. --- # Browser control > Low-level control of the stealth browser plus the interactive browser_act tool. --- # Scrape & research > Read pages the stealth browser can reach, export tabs to files, and run background research reports. --- # Screen > Capture and actuate the Bureau host's own screen — the see→act loop for when the DOM isn't enough. macOS only. --- # Sessions & credentials > Inspect saved identities and open tabs, refresh or re-auth sessions, and store sign-in secrets off the LLM path. --- # Social > Capture footprints, search, read feeds and messages, and take gated write actions on social platforms — as a saved identity, over each platform's own API. --- # Workflows > Discover and run declarative WORKFLOW.md automations as MCP tools. --- # Pricing & credits > What's free (local browsing) versus metered (AI actions), and how credits are debited. --- # Quickstart > Start the daemon and connect it to Claude Code or any MCP client in a couple of minutes. --- # Saved sessions > Save browser identities, sync them from Chrome, and re-authenticate when a login expires. --- # Workflows > List and run declarative WORKFLOW.md automations from the registry — from the CLI or as MCP tools. ---