# Mortakaz Product Discovery and Ghaymah Integration Proposals **Source platform:** [mortakaz.com](https://www.mortakaz.com) — an Arabic platform for discovering technical products and creative ideas, currently listing approximately 686 Arab-built projects. **Target platforms:** [ghaymah.systems](https://ghaymah.systems) (Arabic cloud infrastructure platform) and [mithal.space](https://mithal.space) (Arabic search and indexing engine, built on ghaymah's infrastructure). **Document owner:** _[fill in]_ **Last reviewed:** _[fill in date]_ ## Methodology The products below were identified by browsing the live Mortakaz homepage listing (most recently published projects) and reviewing each product's individual listing page for a working description and live product URL. Two products were selected on the basis of a genuine, identifiable technical fit with ghaymah.systems or mithal.space, rather than generic applicability. Two additional products were noted as close alternatives but not carried forward into full proposals: MilkStraw AI (a cloud cost-optimization advisor, which would integrate with ghaymah as a customer-facing FinOps add-on rather than a hosted workload) and DeployBase (a deployment platform that substantially overlaps with ghaymah's own core offering and is a closer competitor than an integration candidate). --- ## Proposal 1: Bilya Smart Assistant (بِلْيَة المساعد الذكي) **Live product:** [blyhagent.com](https://blyhagent.com) **Maker:** Hossam Youssof ### Product Description Bilya is a professional AI assistant built for car maintenance centers. It handles customer service conversations, provides technical consultations, and allows customers to book maintenance appointments directly through the assistant. It is positioned as a corporate-solutions product for a specific vertical (automotive service businesses) rather than a general-purpose chatbot. ### Integration with ghaymah.systems Bilya's core dependency is an AI inference and knowledge-retrieval stack: a language model for conversation, a way to ground answers in each maintenance center's own service catalog and pricing, and a transactional database for bookings. ghaymah.systems already publishes exactly this stack as first-party products: - **Ready-to-use LLM models** to power the Arabic conversational and consultation layer, replacing or supplementing whatever external model provider Bilya currently calls. - **Vector store and embedding models** to index each maintenance center's service catalog, FAQs, and price lists, enabling retrieval-augmented answers instead of static scripted responses. - **Managed PostgreSQL** ($5/month tier) to store customer records, appointment calendars, and conversation logs. - **Container deployment** for the booking/CRM backend itself, with Git-based automatic deployment. ### Architecture Sketch ``` Car Center Staff End Customer / Car Owner | | v v +----------------------------------------------------------+ | Bilya Chat / Web Front End | | (blyhagent.com) | +---------------------------|--------------------------------+ | API calls v +----------------------------------------------------------+ | ghaymah.systems | | | | +------------------+ +---------------------------+ | | | Ready LLM Model |<---->| Vector Store / Embeddings | | | | (Arabic chat, | | (service catalog, FAQs, | | | | consultation) | | price lists per center) | | | +--------|---------+ +---------------------------+ | | v | | +----------------------------------------------------+ | | | Managed PostgreSQL (bookings, customer records, | | | | appointment calendar) | | | +----------------------------------------------------+ | +----------------------------------------------------------+ ``` ### Added Value for the End User - Faster, more accurate responses because answers are grounded in each center's actual catalog through the vector store, rather than generic model knowledge. - Local-currency billing (Egyptian pound, Saudi riyal, and others, as advertised on ghaymah's pricing page) removes a real friction point for a B2B product selling into small and mid-sized regional automotive businesses that may not hold international payment cards. - Regional data residency for customer contact details and booking history, which matters for a product handling personal data at scale across many independent business customers. - Lower and more predictable infrastructure cost than metered calls to a large international model provider, given ghaymah's published low entry-tier pricing. ### Potential Challenges - **Model quality risk:** ghaymah's public product pages do not disclose which base models power its "ready-to-use LLM models" offering, nor benchmark data against the frontier providers Bilya may currently use. This must be validated with a real evaluation before committing, particularly for consultation-quality technical answers about vehicle maintenance. - **Feature parity risk:** real-time chat with appointment booking typically requires reliable function-calling or tool-use support from the model. This is not documented publicly for ghaymah's LLM product and needs direct confirmation from ghaymah's product team. - **Migration effort:** re-pointing an already-live product's inference and retrieval calls carries downtime and regression-testing risk during cutover. - **Commercial readiness:** Bilya is an early-stage listing on Mortakaz with no recorded community support yet, which may mean limited current infrastructure budget to justify a migration in the near term. --- ## Proposal 2: Hafizah (حافظة) **Live product:** [hafizah.cloud](https://hafizah.cloud) **Maker:** Abdullah Emad ### Product Description Hafizah is a "save everything" system: users save items from their phone, organize them from their computer, and rediscover them from their browser, with full synchronization of their digital library across devices. The product's core promise is captured in its own description: an integrated system so the user never loses anything, wherever they are and whatever device they are using. ### Integration with ghaymah.systems and mithal.space Hafizah has two distinct infrastructure needs that map to ghaymah's two different platforms: - **Storage backend (ghaymah.systems):** the raw saved items (files, images, PDFs, links, notes) need durable, synchronized, multi-device storage. ghaymah's block and object storage, billed transparently at approximately $0.10/GB/month, is a direct fit as the underlying sync backend. - **Search backend (mithal.space):** the product's stated value is not just storage but rediscovery — finding a previously saved item later. mithal.space is described as an Arabic-language search and indexing engine already operating at scale (on the order of billions of indexed pages) on ghaymah's infrastructure. Exposing a private, per-user indexing and full-text search API from mithal.space would let Hafizah offer genuine Arabic-language full-text search over a user's saved library, rather than a simple tag or folder browser. ### Architecture Sketch ``` User (phone / computer / browser) | v +---------------------------------------+ | Hafizah App (hafizah.cloud) | | save / organize / sync client layer | +-------------------|---------------------+ | ----------------------------------- | | v v +-------------------------+ +--------------------------------+ | ghaymah.systems | | mithal.space | | Block / Object Storage | | Private per-user search index | | (saved files, images, |---->| (Arabic full-text indexing, | | PDFs, documents) | index isolated from the public web | | | events index) | +-------------------------+ +--------------------------------+ | v Rediscover: Arabic full-text search results returned to user ``` ### Added Value for the End User - The "rediscover" moment that is the product's central promise becomes materially better: fast, relevant Arabic-language search rather than manual scrolling through folders. - Transparent, low, pay-for-what-you-use storage pricing suits a product whose library size grows continuously and unpredictably per user. - Consolidating both storage and search onto a single regional provider reduces cross-border data-transfer dependencies and aligns with the same local-billing and data-residency advantages described in Proposal 1. ### Potential Challenges - **Product gap on the mithal.space side:** mithal.space is presented publicly as a general web search and indexing engine, not as an exposed private, per-user, multi-tenant search API. Building and securing that API is a real, non-trivial product effort on ghaymah's side before this integration is possible, not a simple configuration change. - **Index isolation:** a private user's saved content must never leak into mithal.space's public web index or become visible to other users. This is an architecturally significant requirement, not a minor detail, and would need a clear multi-tenancy and access-control design. - **Cost profile at scale:** many users saving large numbers of small files (screenshots, short notes, images) may behave differently, cost-wise, than the large-object workloads block storage is typically priced for; this should be modeled before committing. - **Commercial readiness:** like Bilya, Hafizah is an early-stage listing with no recorded community support yet, so real usage volume to justify the integration effort is currently unproven. --- ## Which Integration Is More Viable **Bilya is the more feasible integration in the near term.** The reasoning: 1. **No new product required on ghaymah's side.** ghaymah already publishes ready-to-use LLM models, a vector store, and embeddings as live products. The Bilya integration is primarily a matter of re-pointing existing API calls to an already-shipped product line. The Hafizah integration depends on mithal.space exposing a private, per-user search API that does not appear to exist today — that is new product development, not integration work. 2. **Lower architectural risk.** Bilya's integration risk is mostly about model quality and feature parity, both of which can be validated with a benchmark before committing. Hafizah's integration risk includes a harder problem — guaranteeing strict isolation between a public web search index and private user data — which is a security and architecture concern, not just a quality question. 3. **Clearer commercial fit.** Bilya sells to small and mid-sized regional businesses (car maintenance centers), which is precisely the customer profile ghaymah's own pricing and local-payment positioning is built around. Hafizah is a consumer product, where the local-currency billing advantage is smaller per user and the storage-cost model is less predictable. This assessment is based on what is publicly documented today on mortakaz.com and ghaymah.systems. It should be validated with ghaymah's own product team before any commitment — in particular, the actual language quality of ghaymah's ready LLM models, whether function-calling is supported, and whether a private search API for mithal.space is on ghaymah's roadmap.