# Lumora > Lumora is a web agency that publishes Lumora MCP, a public design-asset and implementation catalog for Codex, other coding agents, and human designers. Lumora MCP is an unrestricted content library. Its metadata helps with discovery and art direction; Codex controls all selections, counts, combinations, placements, source mixes, layers, and repetition. ## Recommended entry points - [Lumora MCP complete HTML context](https://lumoraofficial.de/mcp/context/): Entire manifest and instructions in one static, no-JavaScript HTML page. This is the easiest browser-agent entry point. - [Lumora MCP Codex Reader](https://lumoraofficial.de/mcp/browser/): HTML-compatible searchable record viewer for exact IDs and metadata queries. - [Lumora MCP visual catalog](https://lumoraofficial.de/mcp/): Human-facing previews and catalog filters. - [Manifest](https://lumoraofficial.de/mcp/manifest.json): Canonical machine-readable endpoint map, totals, schemas, and selection policy. - [Codex instructions](https://lumoraofficial.de/mcp/instructions.md): Complete selection, implementation, and rights protocol. - [Full agent context](https://lumoraofficial.de/llms-full.txt): Manifest and instructions combined into one text resource. ## Catalog data - [3D models](https://lumoraofficial.de/mcp/models.json): 1,021 model records with public asset URLs and art-direction metadata. - [Component discovery index](https://lumoraofficial.de/mcp/components-index.json): 1,425 lightweight searchable component records. - [Complete component records](https://lumoraofficial.de/mcp/components.json): Full implementation briefs, accessibility, fallbacks, source links, and performance guidance. - [Images and UI](https://lumoraofficial.de/mcp/image-assets.json): 4,315 image and UI records. - [Animated backgrounds](https://lumoraofficial.de/mcp/animated-backgrounds.json): 152 externally hosted background records. - [Provenance](https://lumoraofficial.de/mcp/provenance.json): Collection origins, rights notes, and catalog provenance. ## Access guidance - Browser and text-only tools should start with [the complete HTML context](https://lumoraofficial.de/mcp/context/), then use [the Codex Reader](https://lumoraofficial.de/mcp/browser/) to search or retrieve individual records. - Direct HTTP clients should use the raw JSON and Markdown endpoints from the manifest. - To open one record in HTML, use https://lumoraofficial.de/mcp/browser/?source=components&id=reactbits-balatro. - To search metadata in HTML, use https://lumoraofficial.de/mcp/browser/?source=components&q=fashion. ## Optional - [Lumora agency website](https://lumoraofficial.de/): Company website and contact information.