Client setup

Your key works with OpenAI- and Anthropic-compatible clients. Point them at the base URL below and you are done.

Try your key right here

Sends the real thing from your browser — the standard POST /v1/chat/completions (and friends) straight to https://api.codezdev-shop.uk. Same wire shape curl, Claude Code and the SDKs send.

Base URL & auth

https://api.codezdev-shop.uk

Every request carries your key as a bearer header:

Authorization: Bearer sk-your-key

curl — OpenAI-compatible

curl https://api.codezdev-shop.uk/v1/chat/completions \
  -H "Authorization: Bearer $KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-sonnet-5","messages":[{"role":"user","content":"Hi"}]}'

Set KEY=sk-your-key first, or paste your key over $KEY.

Claude Code

export ANTHROPIC_BASE_URL=https://api.codezdev-shop.uk
export ANTHROPIC_AUTH_TOKEN=sk-your-key

On older setups the token variable is named ANTHROPIC_API_KEY — either name works with the same value.

Codex & OpenAI SDKs

export OPENAI_BASE_URL=https://api.codezdev-shop.uk/v1
export OPENAI_API_KEY=sk-your-key

Or set the client directly: OpenAI(base_url="https://api.codezdev-shop.uk/v1", api_key="…").

Anthropic API

/v1/messages works with the same bearer key:

curl https://api.codezdev-shop.uk/v1/messages \
  -H "Authorization: Bearer $KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-sonnet-5","max_tokens":64,"messages":[{"role":"user","content":"Hi"}]}'

Image generation

Generate images with gpt-image-2, gpt-image-2.5 or gemini-3.1-flash-lite-image — same key, same base URL:

curl -s -X POST "https://api.codezdev-shop.uk/v1/images/generations" \
  -H "Authorization: Bearer $KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "A fluffy cat",
    "size": "1024x1024",
    "response_format": "b64_json"
  }' \
  | jq -r '.data[0].b64_json' \
  | base64 -d > image.png

The response carries data[0].b64_json (the image bytes, shown decoded above) and data[0].url (a hosted copy) plus revised_prompt. size is optional — it defaults to the model's native square.

Image edit

Send an existing image and describe the change — multipart upload to the edits endpoint:

curl -s -X POST "https://api.codezdev-shop.uk/v1/images/edits" \
  -H "Authorization: Bearer $KEY" \
  -F "model=gpt-image-2" \
  -F "prompt=Make the cat wear a tiny wizard hat" \
  -F "[email protected]" \
  | jq -r '.data[0].b64_json' \
  | base64 -d > edited.png

PNG or JPG input; the reply has the same data[0].b64_json / url shape as generation.

Credits

$1 USDT = $25 of usage at list prices — see Models & pricing.