Clipia AI Gateway

One API for Claude, GPT, Gemini, DeepSeek — pay as you go

Flagship Western and Chinese models on one endpoint. One key, one OpenAI-compatible contract: change the base_url and your OpenAI SDK or langchain code keeps working.

Get a keyDocumentation
Claude · GPT · Gemini
plus Grok, DeepSeek on one key
OpenAI-compatible
change one line — the base_url
Pay as you go
usage-based, billed in credits
Quickstart

Change the base_url — that's it

The same OpenAI-compatible contract. Point it at Clipia and add your key — the rest of the code stays the same.

One parameter — the Clipia AI Gateway address
base_url = "https://api.clipia.ai/v1"
from openai import OpenAI

# The only change vs. a normal OpenAI client: base_url + a Clipia key
client = OpenAI(
    base_url="https://api.clipia.ai/v1",
    api_key="clipia_live_xxxxxxxxxxxxxxxxxxxxxx",
)

resp = client.chat.completions.create(
    model="claude-opus-5",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)

The key starts with clipia_ and is passed like any OpenAI API key. A clipia_test_ key responds instantly and does not spend credits.

Models and pricing

Flagship and standard models

Price per 1M tokens. Pay as you go — the final cost is metered by the actual tokens in the response.

Flagship models

Claude

Opus 5

Output$25
Input$5
per 1M tokens
GPT

GPT-5.5

Output$30
Input$5
per 1M tokens
Claude

Sonnet 4.6

Output$15
Input$3
per 1M tokens
Gemini

3.6 Flash

Output$7.50
Input$1.50
per 1M tokens
Grok

4.3

Output$3.25
Input$1.63
per 1M tokens
DeepSeek

V4 Pro

Output$1.13
Input$0.57
per 1M tokens

Standard models

Gemini

2.5 Flash

Output$3.25
Input$0.39
per 1M tokens
DeepSeek

V4 Flash

Output$0.23
Input$0.12
per 1M tokens
GPT

GPT-5 nano

Output$0.52
Input$0.06
per 1M tokens

Model deep dives: Claude Opus 5 · Gemini 3.6 Flash

Prices are shown per 1M tokens in USD and billed in Clipia credits. Pricing is reviewed manually; live prices are in the /v1/models catalog.

Why Clipia

Every flagship LLM behind one endpoint

The same OpenAI-compatible contract you already use — one key, one integration and transparent credit-based pricing across every flagship model.

Pay as you go

Top up a credit balance and spend it across every model. No seats, no monthly minimums, no lock-in — you pay only for the tokens you use.

One OpenAI-compatible API

A single contract for every model. Works with the OpenAI SDK, langchain, aider and any client that speaks the OpenAI API.

Text, images and video on one key

The same key and the same credit balance cover both LLMs and Clipia image and video generation. No separate Gateway key needed — one integration for every task.

MCP for Claude Code and agents

Plug the models in as an MCP server in Claude Code, Cursor and other agents — no extra glue code required.

Transparent pricing

Pay for the tokens you actually use. Every response returns usage, and the /v1/models catalog shows prices up front.

Western and Chinese flagships together

Claude, GPT, Gemini, Grok and DeepSeek behind a single base_url. Switch models by changing one string — no new account, no new integration.

Frequently asked questions

It is one OpenAI-compatible API for flagship LLMs — Claude, GPT, Gemini, Grok, DeepSeek and more — with transparent credit-based pricing. Effectively a drop-in for apps built on the OpenAI SDK: different models, one contract.

No. Just set the base_url to https://api.clipia.ai/v1 and use a Clipia key instead of the old one. The rest of the code — chat.completions calls, streaming, tools — stays the same.

Top up a credit balance and pay as you go. Billing is metered by the tokens you actually use — every response returns its usage, so there are no seats and no monthly minimums.

Flagships Claude Opus 5 and Sonnet 4.6, GPT-5.5, Gemini 3.6 Flash, Grok 4.3, DeepSeek V4 Pro, plus cheaper standard models. The same key also runs image and video generation.

Yes. Any client with an OpenAI-compatible contract works: langchain, aider, cline and others. Just point the base_url at Clipia and add your key. The models are also available as an MCP server.

One key and one contract replace separate accounts, SDKs and invoices for Claude, GPT, Gemini, Grok and DeepSeek. You get Western and Chinese flagships on one endpoint, a shared credit balance across text, image and video, and transparent per-token pricing — all through the OpenAI-compatible API you already use.

No. One and the same Clipia key works both for LLM chat (this Gateway) and for image and video generation — from a single shared credit balance. If you already have a generation key, you don't need a separate Gateway key: it works right away.

Next

Start with one line of code

Get a key

Create an API key in the developer console — live or test — and drop it into the Clipia base_url.

Get a key

Documentation

Full reference for endpoints, models, streaming and errors — with examples in multiple languages.

Open the docs

For business

Teams, volume tiers, invoicing and dedicated support for organizations.

Go to section