Veriprompt Docs

Appearance

API Quick Start ​

Get a successful API response in under 5 minutes.

Prerequisites ​

  • A VeriPrompt account
  • At least one AI provider configured in Settings > Provider Models (OpenAI, Anthropic, Google, etc.)
  • A VeriPrompt API key (see Step 1 below)

Important: VeriPrompt API Keys vs. Provider Keys

VeriPrompt does not accept raw provider API keys (e.g., sk-ant-... from Anthropic or sk-... from OpenAI) in the Authorization header. You must use a VeriPrompt gateway key (sk-vp-... or vp-gw_sk_...).

Your provider API keys are configured separately in the VeriPrompt dashboard under Settings > Provider Models — VeriPrompt uses them internally to route requests to the correct provider on your behalf.

1. Create a VeriPrompt API Key ​

  1. Log in to the VeriPrompt dashboard.
  2. Navigate to Settings > External API Users.
  3. Click Create Key and select the scope (e.g., Execute Only for chat completions).
  4. Copy the key immediately — it is only shown once. The key starts with sk-vp-.
bash
export VERIPROMPT_API_KEY="sk-vp-your-key-here"

2. Send a Chat Completion (OpenAI-Compatible) ​

VeriPrompt acts as an AI gateway — point any OpenAI-compatible SDK at VeriPrompt's base URL and use your VeriPrompt API key. VeriPrompt routes your request to the best available provider based on your configured routing policies:

bash
curl -X POST https://app.veriprompt.tech/api/v1/chat/completions \
  -H "Authorization: Bearer $VERIPROMPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {"role": "user", "content": "Write a two-sentence summary of AI routing."}
    ],
    "max_tokens": 200
  }'

3. Check the Response ​

A successful response looks like:

json
{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "choices": [{
    "message": {
      "role": "assistant",
      "content": "AI routing dynamically selects the best model..."
    }
  }],
  "usage": {
    "prompt_tokens": 15,
    "completion_tokens": 42,
    "total_tokens": 57
  }
}

4. Use with OpenAI SDKs ​

Python ​

python
from openai import OpenAI

client = OpenAI(
    api_key="sk-vp-your-key-here",
    base_url="https://app.veriprompt.tech/api/v1"
)

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

Node.js / TypeScript ​

typescript
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'sk-vp-your-key-here',
  baseURL: 'https://app.veriprompt.tech/api/v1',
});

const response = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(response.choices[0].message.content);

5. Execute a Stored Prompt ​

If you have prompts stored in the VeriPrompt dashboard, execute them via the MCP endpoint:

bash
curl -X POST https://app.veriprompt.tech/api/mcp \
  -H "Authorization: Bearer $VERIPROMPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "execute_stored_prompt",
      "arguments": {
        "promptId": "prompt_123_abc"
      }
    },
    "id": 1
  }'

Troubleshooting ​

"Invalid API key format" Error ​

json
{"error":{"message":"Invalid API key format. Expected prefix: sk-vp- or vp-gw_sk_","type":"authentication_error","code":"invalid_api_key"}}

Cause: You are sending a raw provider API key (e.g., an OpenAI sk-... or Anthropic sk-ant-... key) instead of a VeriPrompt gateway key.

Fix: Create a VeriPrompt API key in Settings > External API Users. Use the resulting sk-vp-... key in the Authorization header. Your provider keys should be configured separately in Settings > Provider Models.

"Model not available" Error ​

json
{"error":{"message":"Model 'openai/gpt-4o' is not available.","type":"not_found_error","code":"model_not_found"}}

Cause: The specified model is not configured for your company.

Fix: Go to Settings > Provider Models and ensure the model is imported and active. Alternatively, use model: "auto" to let VeriPrompt pick the best available model.

"No active AI providers available" Error ​

Cause: No provider models are configured, or all provider API keys have decryption issues.

Fix: Check Settings > Provider Models and ensure at least one provider is configured with a valid API key.

Next Steps ​