Quick start
Create a key and make your first request to Claude, GPT, Gemini or Grok through one endpoint, with the OpenAI or Anthropic SDK you already use.
Prerequisites
- 1Sign in to Token Share.
- 2Open the API Keys page.
- 3Click Create key, choose the Inference type, and copy the key that starts with
sk-token-. The full key is shown once. - 4Top up with USDT on the Billing page.
- 5Install an SDK if you want one:
pip install openai anthropicornpm install openai @anthropic-ai/sdk.
Token Share speaks two protocol families: OpenAI-compatible (/v1/chat/completions and /v1/responses) and Anthropic (/v1/messages). Which one a model answers on is fixed by the model and shown on its model page. One key works for all of them.
Choose your protocol
OpenAI Chat Completions — the shape most tools support. Grok and Gemini models answer here.
curl https://token-share.app/v1/chat/completions \
-H "Authorization: Bearer $TOKEN_SHARE_KEY" \
-H "content-type: application/json" \
-d '{
"model": "grok-4.6",
"messages": [{"role": "user", "content": "Say hello in one line."}]
}'See what the request cost
Every response carries an x-token-share-request-id header. Pass it to the Billing API with the same key to get that request’s cost and token usage.
# 1. Make a request and keep the id from the response headers
curl -sD - -o /dev/null https://token-share.app/v1/chat/completions \
-H "Authorization: Bearer $TOKEN_SHARE_KEY" \
-H "content-type: application/json" \
-d '{"model": "grok-4.6", "messages": [{"role": "user", "content": "hi"}]}' \
| grep -i x-token-share-request-id
# 2. Look it up
curl https://token-share.app/v1/billing/requests/<request-id> \
-H "Authorization: Bearer $TOKEN_SHARE_KEY"The record is written just after the response finishes. A lookup made the instant a response ends can return 404; retry after a few seconds.