Glossary — MCP, Brand Extraction, Multi-Tenant Documents, and More

MCP — Model Context Protocol: MCP (Model Context Protocol) is an open standard that enables AI agents to use external tools, read data from external sources, and take actions in external systems. Created by Anthropic in late 2024, MCP has since been adopted as a universal standard for AI tool integration.; Before MCP, integrating an external service with an AI model required building a custom integration for each AI provider — one integration for Claude, another for ChatGPT, another for LangChain. MCP unifies this: any MCP server (like DocRocket) can be connected to any MCP-compatible AI without custom code.; MCP is supported by: Claude (claude.ai, Claude API, Claude Desktop), ChatGPT (Pro/Business), Cursor, Gemini, LangChain, LlamaIndex, CrewAI, AutoGen, Semantic Kernel, and most major agent frameworks. DocRocket exposes all its document generation capabilities as MCP tools, enabling any of these agents to generate branded documents without a single line of integration code.; DocRocket's MCP endpoint: https://mcp.docrocket.ai/v1/mcp-tools/docrocket-control/mcp
Brand extraction: Brand extraction is the process of automatically identifying and extracting a company's visual brand identity from their public website URL. DocRocket performs brand extraction by crawling the target URL and algorithmically identifying:; Logo: the primary logo image asset

Brand colors: primary and secondary colors as hex codes

Typography: heading and body font families

Company name: as it appears in the site's metadata; The result is a reusable brand profile stored in your DocRocket workspace. Call docrocket_brand_from_url("https://customer.com") once per customer; all subsequent documents for that customer automatically use the extracted brand. The process takes a few seconds. Brands are fully editable in the web app after extraction.; Traditional document tools require uploading a logo file, entering hex codes, and specifying font names manually — once per customer. Brand extraction eliminates this entirely.
Multi-tenant brand isolation: Multi-tenant brand isolation is an architectural pattern where each customer's brand identity is stored and applied in complete isolation from all other customers' brands. In a multi-tenant document system with brand isolation, generating a document for Customer A will always use Customer A's logo and colors, with no possibility of Customer B's assets appearing or leaking across.; DocRocket enforces brand isolation at the service level. Each brand profile has a unique brand_id. When your AI agent calls docrocket_generate_document, it passes a specific brand_id; DocRocket renders the document exclusively with that brand's assets.; This is especially important for: SaaS products serving multiple customers, agencies managing multiple client brands, and CS teams generating documents for different accounts. DocRocket scales from 1 brand to hundreds — Starter plan supports 3 brands, Scale supports 30, Enterprise is unlimited.
Custom component: A custom component is a reusable document block defined once in the DocRocket web app. Components can contain structured content with design: a pricing table, a signature section, a legal disclaimer, a boilerplate introduction paragraph, a terms and conditions block.; An AI agent references a component by name within document content: include: pricing_table. DocRocket inserts the fully rendered component block at that position in the document. Components automatically inherit the current document's brand — a pricing table defined once will render correctly with Acme Corp's colors on Acme's documents and with GreenLeaf Inc's colors on GreenLeaf's documents.; Components eliminate the need to define common document sections repeatedly in agent prompts. Define once, reference everywhere. Use docrocket_list_components to see all saved components.
Document template: A document template defines the structure, layout, and section arrangement of a document type. DocRocket includes 20+ professionally designed templates: proposals, reports, contracts, invoices, QBRs, case studies, one-pagers, executive briefings, health reports, renewal decks, and more.; When your AI calls docrocket_generate_document, it specifies a template (e.g., template: "proposal"). DocRocket renders that template structure with the specified brand applied — adapting colors, logo placement, and typography automatically. Use docrocket_list_templates to see all available templates with required content fields.
OAuth (as used by DocRocket): OAuth (Open Authorization) is an open standard for authorization that lets a user grant an application access to their data in another application without sharing their password. DocRocket uses OAuth for all MCP connections — no API keys to create, copy, or rotate.; When you add DocRocket as a connector in Claude, ChatGPT, or Cursor, the AI prompts you to sign in to your DocRocket account. This is the OAuth flow. After signing in once, the connection is established. Multiple AI agents signed in with the same DocRocket account share the same workspace: brands, components, templates, and document quota.
Brand ID: A brand ID is the unique identifier for a customer brand profile in your DocRocket workspace. After calling docrocket_brand_from_url, a brand ID is returned (e.g., br_abc123). This ID is passed to docrocket_generate_document to apply that brand to a document.; Best practice: store the brand ID in your own database alongside the customer record the first time you extract a brand. On subsequent document generation calls for that customer, pass the stored brand ID directly — skipping the extraction step entirely (saving time and reducing external crawls).
PDF generation: PDF (Portable Document Format) generation is the process of converting a document into a PDF file — a format that preserves layout, fonts, and visual design across all devices and operating systems. DocRocket generates a PDF for every document automatically, alongside a shareable web URL. Both are available within 2 seconds.; DocRocket uses a managed headless browser rendering pipeline to produce PDFs. This eliminates the need for developers to manage Puppeteer, headless Chrome, or other rendering infrastructure. See DocRocket vs Building Your Own for a full comparison.