Learn how to use the vault

AI Khazna is a curated vault of AI assets for the Arabic-speaking world and beyond. It gathers skills, MCP servers, agents, prompts, and workflows from across the ecosystem, vets them, labels them for safety, and organizes them by category and platform so you can find what works and skip what does not.

Yes. Browsing the vault, reading guides, joining the community, and using the free assets and bundles costs nothing. Some premium features and bundles may be offered separately.

Assets are the building blocks you add to an AI assistant. AI Khazna covers five main types: skills (extend what an assistant can do), MCP servers (connect it to external tools and data), agents (ready-to-run templates), prompts (ready-made instructions), and workflows (multi-step chains).

Each asset is tagged with the platforms it works on, including Claude, ChatGPT, Gemini, Cursor, and Lovable, among other major AI agents. You can filter and browse by the platform you use.

Open the asset page, confirm your platform is listed as compatible, then follow the install steps shown — usually copying an install command, adding a connector, or pasting a prompt. The Help Center has step-by-step guides for each asset type and platform.

Green means we reviewed the asset and found no obvious concerns (not a guarantee). Yellow means the asset is usable but you should review it yourself before installing. The public catalog only shows green and yellow items; unverified items are never shown publicly.

Yes. AI Khazna is fully bilingual (English and Arabic) with right-to-left support. Every asset, guide, bundle, and community space is available in both languages.

Bundles are hand-picked sets of assets organized around a goal and a skill level — beginner, intermediate, or advanced — so you can set up a complete workflow (for example, a one-person social media studio or a developer accelerator) without assembling it piece by piece.

Send it in with a name, short description, asset type, the platforms it works on, and a source link. The team reviews it for safety, assigns a security label, writes a bilingual description, and publishes it with credit to the author.