The prepaid, metered
OpenRouter alternative
NeoGPT AI is an OpenRouter-style API gateway with prepaid credits, OpenAI-compatible routes, chat playground, API keys, model catalog, usage logs, and billing controls.
curl https://api.neogpt.engineering/v1/chat/completions \
-H "Authorization: Bearer $NEOGPT_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-5.5",
"messages": [
{ "role": "user", "content": "Draft a launch note." }
]
}'- OpenAI
- Anthropic
- Gemini
- Meta
- Mistral
- xAI
- DeepSeek
- Qwen
- Black Forest Labs
- Moonshot
- Cohere
- Perplexity
- Stability AI
Your stack already speaks our API.
Point any OpenAI SDK at one base URL and switch models by name. Chat, code, and image all run off the same balance and the same usage log.
{
"model": "gpt-5.5",
"messages": [
{ "role": "user", "content": "Summarize this PR." }
]
}Metered access to current frontier models.
GPT-5.5
OpenAI
$10/$60 per 1M tokens
From signup to first token in minutes.
Pay as you go credits. No subscription.
Add a balance once and spend it across every model. Drag the slider to estimate monthly metered spend at the current public rate card.
Estimate at a 70/30 input/output mix. Credits sell 1:1, so 1M credits equals $1 of metered API usage.
Transparent rates: every model is metered per token at the rates shown above in Available models. Credit packs sell 1:1, and margin lives in the model rate card rather than hidden package multipliers.
Built to carry real production traffic.
Your keys never leak
You hold one NeoGPT key. The real provider keys stay locked on our servers.
One bill, nothing hidden
Every request logged with model, tokens, latency, and exact cost. One invoice.
No rewrite required
Speaks the OpenAI API. Change one base URL — keep your SDK, prompts, and stack.
One key for all your model traffic.
Buy credits, copy a key, and start sending requests. No setup wizard, no agent workflow, no subscription.