Model inference for African regions · Private beta
Test open-model inference for African workloads.
Start on Yamify-managed US canary capacity. Move to a verified African region only when the selected endpoint shows it as ready.
Founder-reviewed · 100 test credits for approved builders · No card required
How it works
From API key to the first completion.
Use the same API shape developers already know. Yamify handles model deployment, routing, scaling, and the regional infrastructure underneath.
- 01
Request access
Tell us your product, model, expected traffic, latency target, and preferred African region.
- 02
Choose a ready model
Use only models marked ready for your account and review the actual serving location.
- 03
Call one API
Use an OpenAI-compatible SDK and point it at the Yamify endpoint.
- 04
Verify and scale
Track usage, serving location, and performance while Yamify manages the GPUs.
Clear terms, no surprises
Who qualifies
- Open to founders, developers, operators, students, and teams
- A work email and concrete project description are required
- One free-compute grant per approved builder or company
- Requests are reviewed manually and answered within 24 hours
- Access is subject to capacity, safety, and acceptable-use review
Questions
Early access FAQ
Who can request Yamify early access?
Anyone building a real product, workflow, or business can request access. Every request is reviewed by a Yamify founder, with priority given to teams ready to test a concrete workload.
What is included?
Approved builders receive 100 Yamify credits for the inference private beta and access to Yamify's application workspace. No card is required to start.
Are AI model costs included?
The private beta focuses on Qwen, Mistral, and Gemma through one OpenAI-compatible API. Premium third-party model usage is separate.
Does my data stay in Africa?
Not yet as a production guarantee. Yamify's current production runtime is in the United States while the African inference data plane is being built. Every endpoint will show its actual serving region; do not send residency-regulated data until the endpoint carries verified African region evidence.
What happens when my Yamify credits run out?
When credits reach zero, paid inference stops. Workspace data stays preserved, and the builder can resume after adding credits.
Yamify founding access
Test one real workload.
Tell us the model, traffic, latency, and residency you need. We will approve access only when a suitable target is available.