cURL Integration

cURL is the quickest way to test VaultProxy. All examples below can be pasted directly into your terminal.

Basic Request

curl -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer vpx_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {"role": "user", "content": "What is the capital of Poland?"}
    ]
  }'

Using Different Models

Anthropic Claude

curl -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer vpx_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-4.6-sonnet",
    "messages": [
      {"role": "user", "content": "Explain quantum computing in simple terms."}
    ]
  }'

Google Gemini

curl -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer vpx_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-2.5-pro",
    "messages": [
      {"role": "user", "content": "Hello from Gemini!"}
    ]
  }'

Bielik (Polish Model)

curl -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer vpx_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "bielik/bielik-13b-pro",
    "messages": [
      {"role": "user", "content": "Opowiedz mi o historii Krakowa."}
    ]
  }'

System Messages

curl -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer vpx_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {"role": "system", "content": "You are a Polish legal expert. Answer in Polish."},
      {"role": "user", "content": "Czym jest RODO?"}
    ],
    "temperature": 0.5,
    "max_tokens": 1000
  }'

Inspecting the Anonymized Prompt

Add the X-Show-Raw-Prompt: true header to see exactly what VaultProxy sends to the AI provider after PII anonymization. The anonymization details are returned in the X-Anonymization-Info response header as base64-encoded JSON.

curl -v -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer vpx_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Show-Raw-Prompt: true" \
  -d '{
    "model": "google/gemini-2.5-flash",
    "messages": [
      {"role": "user", "content": "Summarize this: Jan Kowalski, PESEL 85031501234, email [email protected], phone +48 600 123 456"}
    ]
  }'

The response body is a normal OpenAI-compatible response (with de-anonymized data). The response header X-Anonymization-Info contains the base64-encoded anonymization details:

# Decode the header to see what AI actually received:
echo "BASE64_VALUE_FROM_HEADER" | base64 -d | python3 -m json.tool

{
  "anonymized_messages": [
    {
      "role": "user",
      "content": "Summarize this: <PERSON_1>, PESEL <PESEL_1>, email <EMAIL_ADDRESS_1>, phone <PHONE_NUMBER_1>"
    }
  ],
  "raw_llm_response": "<PERSON_1>, with PESEL <PESEL_1>, resides in...",
  "entities_found": 4,
  "entity_types": ["PERSON", "PESEL", "EMAIL_ADDRESS", "PHONE_NUMBER"],
  "mapping": {
    "<PERSON_1>": "***",
    "<PESEL_1>": "***",
    "<EMAIL_ADDRESS_1>": "***",
    "<PHONE_NUMBER_1>": "***"
  }
}

info

The mapping field shows tags but masks actual values with *** for security. The real values are only used internally for de-anonymization and are never exposed in headers.

## Health Check

Verify that VaultProxy is running and reachable:

```bash
curl https://api.vaultproxy.ai/v1/health

Expected response:

{
  "status": "ok",
  "version": "1.0.0"
}

Streaming

curl -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer vpx_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -N \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {"role": "user", "content": "Write a haiku about Warsaw."}
    ],
    "stream": true
  }'

tip

The -N flag disables output buffering so you see tokens as they arrive.

Common Options

Parameter	Type	Description
model	string	Model identifier (required)
messages	array	Conversation messages (required)
temperature	float	Randomness, 0.0 to 2.0 (default: 1.0)
max_tokens	integer	Maximum response tokens
stream	boolean	Enable streaming (default: false)
top_p	float	Nucleus sampling threshold
stop	string/array	Stop sequences

Error Responses

# 401 - Invalid API key
curl -X POST https://api.vaultproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer invalid_key" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-4o", "messages": [{"role": "user", "content": "test"}]}'

{
  "error": {
    "type": "authentication_error",
    "message": "Invalid API key provided."
  }
}