Skip to main content
POST
/
v1
/
chat
/
{chatId}
/
stream
curl -N -X POST https://api.vouchmark.com/v1/chat/chat_aBcD.../stream \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"message": "Summarise my monitoring alerts from last week"}'
Same behaviour as Send message, but the AI reply is streamed back token-by-token for real-time rendering. The user message and the final AI message are still persisted to the chat.
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.

​
Path parameters

​
chatId
string
required
The chat session ID.

​
Body

​
message
string
required
Your message, 1–4000 characters.
curl -N -X POST https://api.vouchmark.com/v1/chat/chat_aBcD.../stream \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"message": "Summarise my monitoring alerts from last week"}'

​
Response format

The response is sent with Content-Type: application/x-ndjson. The body is a sequence of newline-delimited JSON objects, one per line. Each object has the shape { "type": <string>, "data": <any> }. This is NDJSON, not Server-Sent Events, so read it with a streaming body reader rather than EventSource: 
{"type":"text","data":"Based on"}
{"type":"text","data":" your monitoring"}
{"type":"text","data":" data from the past week..."}
{"type":"grounding_sources","data":{"groundingChunks":[{"web":{"uri":"https://...","title":"FIRS"}}]}}
{"type":"finish","data":{"complete":true}}

​
Event types

typedataDescription
textstringA text fragment of the reply. Concatenate all text events for the full message.
tool_call{ toolName, args }The AI invoked a tool (e.g. web search) mid-response. Informational.
grounding_sources{ groundingChunks, searchEntryPoint? }Sources used to ground the answer. Emitted once, before finish, only when present.
finish{ complete: true }The stream is complete.
error{ message }An error occurred mid-stream. The connection then closes.

​
Client example

const response = await fetch(`/v1/chat/${chatId}/stream`, {
  method: "POST",
  headers: {
    Authorization: `Bearer ${token}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ message: "..." }),
});

const reader = response.body!.getReader();
const decoder = new TextDecoder();
let buffer = "";

while (true) {
  const { done, value } = await reader.read();
  if (done) break;

  buffer += decoder.decode(value, { stream: true });
  const lines = buffer.split("\n");
  buffer = lines.pop() ?? "";

  for (const line of lines) {
    if (!line.trim()) continue;
    const event = JSON.parse(line);
    if (event.type === "text") {
      appendToUI(event.data);
    }
  }
}