The AI that does the work for you
Hawiyat AI Composer
Draft the weekly report, reconcile an invoice, or turn a spec into code. It picks the best AI, connects your systems, and ships a result it already checked. Billed in DZD, supported from Algeria.
Already a Composer client? Open your usage dashboardQuestions? Reply in Arabic, French, or English on WhatsAppPartners & early customers
Proof, not promises
Proof it works.
Straight from our own operations, we only print what’s real. Figures verified against the Hawiyat operations dashboard as of August 9, 2026. See how we count.
satisfied clients and businesses
resellers
tokens executed through Composer
Pricing
Plans for the execution layer
One layer, every model, and everything billed in dinars.
Open your usage dashboardHawiyat AI Composer
For solo builders. Give it a task, get a checked result.
- Model-agnostic routing per task
- Context-aware execution against your systems
- Automatic model fallbacks on failure
- Evaluation and quality score for every run
- Semantic caching to cut repeat spend
- Billed in DZD with a transparent per-task cost
No card needed. CCP or Baridi Mob.
Hawiyat AI Composer MAX
5× more tasks at the same time.
For startups and teams shipping every day.
- 5X base execution capacity, more parallel runs and tasks
- Model-agnostic routing per task
- Advanced semantic caching (vector-based)
- Automatic model fallbacks on failure
- Evaluation and quality score for every run
- Billed in DZD with a transparent per-task cost
No card needed. CCP or Baridi Mob.
Enterprise
The execution layer, sized for your entire operation. Custom infrastructure, capacity, and support.
- The execution layer, tuned for your whole operation
- Dedicated account manager who knows your runs and your stack
- Priority WhatsApp support from the local team in AR, FR, and EN
- Custom infrastructure and capacity built around your workloads
- Single contract, one invoice, everything billed in DZD
Social proof
What clients say
Real people, running real work on the Hawiyat execution layer.
Very good, it works really well. Our greenhouse automation just runs, day after day.
I came in as a beginner and support walked me through every step. When I needed more tokens, they topped me up. Support 100/10.
We started with n8n hosting and came back six months later for Composer MAX 5X. When a provider grows with you, you stay.
I had never touched Claude Code before. They explained it step by step until I understood what was actually running.
I did not know n8n at all and had no time to babysit servers. They set it up hosted and ready, so I could get back to client work.
Customer quotes shared with permission. Roles shown, names kept private.
FAQ
Questions, answered straight
Hawiyat takes a task you would normally do by hand and makes it run itself. You describe what you want, like a WhatsApp order confirmed or a support reply sent, and it picks the AI, pulls in the context from your tools, and checks the answer before it goes anywhere. Whatever AI exists tomorrow, you'll need a layer that decides how to use it.
No. Models are routes, chosen per task by quality, latency, and cost. GPT, Claude, Gemini, and open models all sit behind the same layer, so a provider change never breaks your pipeline. The layer outlives any single model.
Everything is billed in Algerian dinars. Composer caches repeated work, routes each task to the most efficient model, and logs a transparent per-task cost, roughly a fraction of a DZD per task. You always see what you paid and what you got, measured, not guessed. Composer plans run 6,000 DA/month (Pro), 15,000 DA/month (MAX 5X), and 30,000 DA/month (MAX 20X).
Yes, everything is billed in DZD. Pay with CCP or Baridi Mob, no foreign card needed.
No. Your data is never used to train models. Runs carry your context to complete the task, and the evaluation logs are yours to audit. Data stays between your systems and the layer.
Pick a Composer plan on the home page, or browse n8n, WhatsApp, and hosting on the services page. We activate your workspace within 24 hours, you connect your tools, and run your first task. A WhatsApp workflow can be executing the same day. Prefer to talk first? Chat with us on WhatsApp and we'll set you up.
A chatbot answers a prompt. Composer executes a business task: it pulls your customer and order context, picks the best model for the job, runs it against your systems, and checks the result before it reaches anyone. Same engines, but the layer decides the how.
Yes, WhatsApp is a first-class route on the layer. Support, sales, and back-office runs happen there, and a workflow can be executing the same day you order.
Ready when you are
Have your first task running today.
Pick a plan, connect your tools, and send your first task. Hawiyat handles the models. You handle the business.