GPT
OpenAI
General-purpose flagship for chat, coding, tool use and structured output.
Antananarivo, Madagascar · Serving Africa
MadAfri Nexus AI
One API key for every AI model
Chat with GPT, Claude, Gemini, DeepSeek, GLM and Kimi. Create images with GPT Image 2.5, Grok Imagine and Nano Banana. Generate video with Seedance, Wan 3.0 and MiniMax H3. One endpoint and one balance, built for developers and businesses across Africa.
- Pay as you go
- OpenAI-compatible
- Text, image and video with one key
Start with one request
Pick a model and a language. The same key and the same base URL work for text, images and video.
curl https://api.madafri.com/v1/chat/completions \
-H "Authorization: Bearer $MADAFRI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "claude-sonnet-5",
"messages": [
{
"role": "user",
"content": "Explain how a mobile money payment API works, in three short sentences."
}
]
}'Example request. Exact model IDs and parameters are listed in your console.
Every major model, one endpoint
Switch models by changing one string. New versions are added as they ship.
Chat & reasoning
Claude
Anthropic
Strong on long documents, careful writing, coding and agent workflows.
Gemini
Multimodal by default: text, images, audio and very long context.
DeepSeek
DeepSeek AI
High-value reasoning and coding models with open weights.
GLM
Zhipu AI · Z.ai
Bilingual general models with solid agent and coding skills at low cost.
Kimi
Moonshot AI
Long-context models tuned for research, coding and agentic tasks.
Image generation
GPT Image 2.5
OpenAI
OpenAI’s newest image model: faster generation, faithful edits and consistent style across a series.
Grok Imagine
xAI
xAI’s image model for quick, vivid drafts and social-ready visuals.
Nano Banana
Google’s Gemini image model, strong at precise edits, layouts and legible text in images.
Built for teams shipping across Africa
Not a model showroom: a shorter path from an idea to a working product.
Developers & startups
Prototype with the best model for each job without opening six provider accounts.
Agencies & content studios
Turn briefs into images and video for campaigns, social content and product pages.
Businesses & enterprises
Add assistants, document processing and translation to the tools your teams already use.
Schools & tech hubs
Give students and cohorts hands-on access to frontier models with one shared setup.
Live in three steps
If your code already talks to the OpenAI API, you are one line away.
Create your account
Sign up in the console and create an API key for each project or environment.
Add credit
Top up your balance and pay only for what you use. Live per-model rates are shown in the console.
Point your code at MadAfri
Set the base URL to api.madafri.com/v1, keep your existing SDK and choose any model.
Video generation, pay as you go
Submit a job, poll the task, download the clip. Per-second rates are shown live in the console.
Seedance 2.5
NewestByteDance
Billed per second · live rate in console
New-generation multimodal generation and editing, for longer takes and richer references.
- Takes up to 30 seconds
- Text, image, video and audio references
- Audio generated in sync with the picture
Seedance 2.0
ByteDance
Billed per second · live rate in console
A dependable workhorse for everyday commercial work, from quick drafts to key shots.
- Text-to-video and image-to-video
- Multi-reference input
- Tiers for speed or quality
Wan 3.0
Alibaba
Billed per second · live rate in console
Long clips from almost any input, including documents and slides.
- Clips up to 30 seconds
- Text, image and document input
- Sound generated with the video
MiniMax H3
MiniMax · Hailuo
Billed per second · live rate in console
One multimodal model for generating, referencing and editing video with sound.
- Up to 2K resolution, 15-second clips
- Native stereo audio
- Text, image, video and audio as context
Drop-in for the tools you already use
MadAfri speaks the OpenAI API format, so SDKs, agent frameworks and editors that accept a custom base URL work out of the box. Change two environment variables and ship.
Read the API docs# .env: the only two lines that change
OPENAI_BASE_URL="https://api.madafri.com/v1"
OPENAI_API_KEY="sk-madafri-..."
# then use any OpenAI-compatible client
from openai import OpenAI
client = OpenAI()
client.chat.completions.create(model="deepseek-v4", messages=[...])Questions, answered
The essentials on access, billing and compatibility.
What is MadAfri Nexus AI?
An all-in-one AI platform from Antananarivo, Madagascar. You get one API key and one balance for text, image and video models from many providers, instead of managing a separate account for each.
Will my existing OpenAI code work?
Yes. The API follows the OpenAI format. Set the base URL to https://api.madafri.com/v1, use your MadAfri key and pick a model. Official OpenAI SDKs and most frameworks that accept a custom base URL work unchanged.
How does billing work?
Pay as you go, with no subscription. Text models are billed by tokens, images per image and video per second. The current rate for every model is shown in the console before you call it.
Which models can I use?
GPT, Claude, Gemini, DeepSeek, GLM and Kimi for chat; GPT Image 2.5, Grok Imagine and Nano Banana for images; Seedance 2.5 and 2.0, Wan 3.0 and MiniMax H3 for video. The console always lists what is live.
How do I top up and get support?
Top-up options are shown in the console when you add credit. For help, message us on our Facebook page and we will reply in English or French.
Are you affiliated with the model providers?
No. MadAfri is an independent gateway. Model names and brands belong to their respective owners, and each provider’s usage policies still apply to your requests.