DEV Community

Daniel Dong
Daniel Dong

Posted on

3 features that made my API gateway actually usable. 1 model that made it worth using.

3 features that made my API gateway actually usable. 1 model that made it worth using.

Feature 1: One-click GitHub login

Registration used to be a form. Now it's a button.

Click GitHub → authorize → dashboard. Five seconds. No email.
No password. No verification code. The registration page and login
page both have it. One click, you're in.

Feature 2: Smart onboarding

After login, instead of dropping you on an empty dashboard, you land
on a page that hands you everything at once:

🎉 Your account is ready!
🔑 API key: mb-xxx [📋 Copy]

curl https://aibridge-api.com/v1/chat/completions
-H "Authorization: Bearer mb-xxx"
-d '{"model":"deepseek-chat",...}' [📋 Copy]

import openai
client = openai.OpenAI(
api_key="mb-xxx",
base_url="https://aibridge-api.com/v1"
) [📋 Copy]
Enter fullscreen mode Exit fullscreen mode

Copy, paste, run. 30 seconds from login to first response.

2

Feature 3: Usage bar that escalates

The dashboard shows how many tokens you've used. But it doesn't just
show a number — it changes behavior as you approach the limit:

0-50%: [████░░░░░░] 42% — small upgrade button, stay calm
50-80%: [████████░░] 68% — full-width button, time to think
80-90%: [█████████░] 84% — orange bar, "Unlock 10x more →"
90-100%:[██████████] 97% — red flashing bar, "15K tokens left!"
Enter fullscreen mode Exit fullscreen mode

You notice it exactly when you need to. Not before. Not after.

The model: Kimi K3

The reason people try the playground in the first place.

K3 has a 1M token context window — 4x Claude, 8x GPT-4o. Drop a
codebase. Get a review. No chunking required.

And it reasons through every request without a toggle. No "think mode
ON/OFF." No prompt engineering. Just send the prompt. K3 thinks, then
answers.

import openai
client = openai.OpenAI(
api_key="mb-xxx",
base_url="https://aibridge-api.com/v1"
)
client.chat.completions.create(
model="kimi-k3",
messages=[{"role":"user","content":"Find every race condition in this module"}]
)
Enter fullscreen mode Exit fullscreen mode

1

Same endpoint as deepseek-chat ($0.27/M), qwen-max (multilingual),
and glm-4-plus (complex reasoning). Same code. Different model string.

The playground

Free. No signup. 10 requests/day. All 15 models including K3.
If you like what you see, registration is one click away.

3

15 Chinese AI models. One OpenAI-compatible endpoint. Free tier:
500K tokens/month. GitHub login in one click.

aibridge-api.com/playground.html (try K3, no signup)
aibridge-api.com/register.html (GitHub login, one click)

5

Top comments (0)