> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vouchmark.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Quick chat

> Single-turn AI query against a company without persisting a chat.

A stateless alternative to the full chat flow. Ask a question scoped to a company and get an AI reply. Nothing is saved — no chat or messages are persisted. You can pass prior turns via `conversationHistory` to keep context across calls yourself.

<Note>
  Requires a positive wallet balance. Billed at ₦100 per 1,000 tokens consumed by the request, and rate limited to 10 AI requests per minute.
</Note>

## Body

<ParamField body="companyId" type="string" required>
  The company to scope the answer to.
</ParamField>

<ParamField body="message" type="string" required>
  Your question, 1–4000 characters.
</ParamField>

<ParamField body="conversationHistory" type="array">
  Prior turns to give the model context. Each item is `{ "sender": "user" | "ai", "content": string }`. Optional.
</ParamField>

<RequestExample>
  ```bash cURL theme={null}
  curl -X POST https://api.vouchmark.com/v1/chat/quick \
    -H "Authorization: Bearer $TOKEN" \
    -H "Content-Type: application/json" \
    -d '{
      "companyId": "cmp_aBcD...",
      "message": "When is the next CAC annual return due?",
      "conversationHistory": [
        { "sender": "user", "content": "Tell me about my compliance status" },
        { "sender": "ai", "content": "Your company has 6 active obligations..." }
      ]
    }'
  ```
</RequestExample>

<ResponseExample>
  ```json theme={null}
  {
    "message": "Response generated successfully",
    "userMessage": {
      "sender": "user",
      "content": "When is the next CAC annual return due?",
      "timestampAt": "2026-06-27T10:00:00Z"
    },
    "aiMessage": {
      "sender": "ai",
      "content": "Your CAC annual return is due by 30 June 2026, in 3 days.",
      "timestampAt": "2026-06-27T10:00:02Z",
      "groundingSources": null
    }
  }
  ```
</ResponseExample>

`userMessage` and `aiMessage` are returned for convenience but are not stored. `groundingSources` is `null` unless the AI used web search.