- OpenAI-compatible API proxy for forwarding requests.
- Cohere reranking through the Azure AI Foundry endpoint.
- Web UI to create and manage API keys (including deactivation).
- Per-key usage tracking with filtering and sorting in the UI.
- Admin usage dashboard with range filters (24h, 7d, 30d, all).
- Admin cost dashboard.
- Release notes page linked from the sidebar.
git clone git@github.com:bat-bs/openai-proxy.git
cd openai-proxy
go build cmd/main.go -
Create your own .env file with the .env.example You can use the following Link as a reference for filling the Values: https://learn.microsoft.com/en-us/azure/ai-services/openai/reference#rest-api-versioning
-
./main -
Generate Key in Web ui
-
Use it:
payload="{\"messages\":[{\"role\":\"system\",\"content\":[{\"type\":\"text\",\"text\":\"You are an AI assistant that helps people find information.\"}]}],\"temperature\":0.7,\"top_p\":0.95,\"max_tokens\":800,\"model\":\"gpt-4o\"}"
curl "http://localhost:8082/api/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer <Key-From-Web-UI>" \
-d "$payload"Reranking uses the same DEPLOYMENT_NAME, BASE_URL, and AZURE_API_KEY
configuration as the Azure OpenAI proxy. BASE_URL is the shared Azure domain
suffix, for example azure.com. The proxy exposes reranking at
/api/v1/rerank:
The Rerank model must be registered in the admin model catalog as type
RERANK. The current pricing policy for Cohere-rerank-v4.0-fast is $2 per 1,000 searches; this is seeded for that model only and is not a general price
for other Cohere models.
curl "http://localhost:8082/api/v1/rerank" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer <Key-From-Web-UI>" \
-d '{
"model": "Cohere-rerank-v4.0-fast",
"query": "What are the health benefits of green tea?",
"documents": [
"Green tea contains antioxidants.",
"Basketball is one of the most popular sports in the United States."
]
}'For Open Tasks i use the Github Issues.