Chat

A multi‑model LLM chat. Conversations persist server‑side, and assistant replies stream back as Server‑Sent Events.

List models

GET /v1/chat/models Public

curl https://ai-api.vanguardinitiative.com/v1/chat/models

200 Available models

{
  "status": "success",
  "models": [
    { "id": "<model id>", "tier": "fast", "label": "Fast", "creditsPerMessage": 1 },
    { "id": "<model id>", "tier": "smart", "label": "Smart", "creditsPerMessage": 2 }
  ]
}

Two tiers: Fast, the default, and Smart, for harder problems and long documents. Wherever you name a model, pass fast or smart — or an id from this list, which may change as we update the models behind each tier. Read tier, not the id. An unknown model returns 400 INVALID_MODEL.

Create a conversation

POST /v1/chat/conversations API key

JSON body

`title` string	Optional conversation title.
`model` string	Model id from /v1/chat/models.
`projectId` uuid	Attach to a project.

Send a message (streaming)

POST /v1/chat/conversations/{id}/messages API key

The response is a text/event-stream of tokens. If the connection drops mid‑reply, reconnect with GET /v1/chat/conversations/{id}/stream (returns 204 when idle).

curl -N -X POST \
  https://ai-api.vanguardinitiative.com/v1/chat/conversations/$CONV_ID/messages \
  -H "authorization: Bearer $API_KEY" \
  -H 'content-type: application/json' \
  -d '{"content":"Summarise this contract in Lao."}'

const res = await fetch(
  `https://ai-api.vanguardinitiative.com/v1/chat/conversations/${convId}/messages`,
  {
    method: 'POST',
    headers: {
      authorization: 'Bearer ' + API_KEY,
      'content-type': 'application/json',
    },
    body: JSON.stringify({ content: 'Summarise this contract in Lao.' }),
  },
);
const reader = res.body.getReader();
const decoder = new TextDecoder();
for (;;) {
  const { value, done } = await reader.read();
  if (done) break;
  process.stdout.write(decoder.decode(value)); // SSE chunks
}