Early access is open
The world's best open AI models, paid for in naira.
No dollar card, no declined payments. Change one line – your base URL – and run DeepSeek, Qwen, Kimi and gpt-oss from your app. OpenAI-compatible, so your existing code just works.
Signups are open. Accounts are approved in batches. Naira top-ups via Paystack are coming soon.
- base_url="https://api.openai.com/v1"
+ base_url="https://api.baseurl.dev/v1"
Same SDK, same request format, open-weight models.
Start in three steps
-
Join early access
Sign up with your email and verify it. We approve accounts in batches and email you when yours is ready.
-
Create an API key
Once approved, create a key in your dashboard and set a monthly budget on it if you like.
-
Change your base URL
Point your OpenAI SDK or HTTP client at
https://api.baseurl.dev/v1and pick a model below.
Built for developers who build here
Everything you need to ship AI features without a foreign card or a new SDK.
-
Pay in naira
Top up your balance through Paystack with your card, a bank transfer or USSD. No dollar card, no FX surprises at checkout.
Coming soon during early access -
Works with your code
The same API as OpenAI. Keep the official OpenAI SDKs, or any library or HTTP client that lets you set a base URL: point it at BaseURL and it just works.
-
Stay in control
Set a budget on every API key, watch usage live in your dashboard, and know that we never log your prompts or completions.
One line to switch
Set the base URL to https://api.baseurl.dev/v1 and use your BaseURL key. Nothing else in your code has to change.
# pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://api.baseurl.dev/v1",
api_key="YOUR_BASEURL_KEY",
)
reply = client.chat.completions.create(
model="deepseek-v4p1-flash",
messages=[{"role": "user", "content": "Hello from Lagos"}],
)
print(reply.choices[0].message.content)
curl https://api.baseurl.dev/v1/chat/completions \
-H "Authorization: Bearer $BASEURL_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v4p1-flash",
"messages": [
{"role": "user", "content": "Hello from Lagos"}
]
}'
Leading open-weight models
Pick the right model for the job. Switch between them by changing the model field.
| Best for | Model | Input | Output |
|---|---|---|---|
| Lowest cost, reasoning | gpt-oss-120b gpt-oss-120b |
$0.20 | $0.70 |
| Fast everyday tasks | DeepSeek V4.1 Flash deepseek-v4p1-flash |
$0.40 | $1.35 |
| Smart all-round | Qwen 3.8 Max qwen-3p8-max |
$2.30 | $6.90 |
| Top quality, coding | Kimi K3 kimi-k3 |
$3.50 | $16.50 |
Frequently asked questions
Is it OpenAI-compatible?
Yes. BaseURL speaks the OpenAI Chat Completions API. Use the official OpenAI SDKs or any tool that lets you set a base URL, set it to https://api.baseurl.dev/v1, and use your BaseURL API key.
Do you store my prompts?
No. We do not log or store your prompts or completions. They pass through to the inference provider that runs the model, only to generate your response. We keep usage counts, such as tokens, model and time, for billing.
Which models can I use?
At launch: gpt-oss-120b, DeepSeek V4.1 Flash, Qwen 3.8 Max and Kimi K3. Your dashboard always shows the current list and prices. We add models as strong new open-weight releases arrive.
When can I pay in naira?
Naira top-ups through Paystack (card, bank transfer and USSD) are coming soon during early access. We will email early-access members as soon as they are live, together with naira pricing.
Who can sign up?
Developers and teams in supported countries, starting with Nigeria. Anyone can join the early-access list today; we approve accounts in batches.
How do I get support?
Email [email protected].
Start building with BaseURL
Join early access now. We approve accounts in batches and will email you when yours is ready.