MONSOON
Tiếng Việt

Monsoon Gateway

Reliable access toleading modelsthrough one API.

One OpenAI-compatible API. Ordered model fallback. A traceable record for every routed request.

View API docsSign in
Vietnam-resident routingComing soon

Request trace

Request IDReturned with every responseTimingGateway latencyRead fallback guideGuide
  1. Primary model01Your first choice
  2. Route failedPrimary route unavailableEligible failure before output
  3. Ordered fallback02Next model in your order
  4. Served modelFallback response returnedDelivered to the caller
  5. Audit recordTrace savedRoute, latency, usage, cost estimate
MissionMonsoon is building the infrastructure and research capacity to develop and deploy frontier AI in Vietnam—starting with one reliable gateway to the world’s leading models.

Reliability

Automatic fallback, in the order you choose.

Choose a primary model and an ordered fallback list. On an eligible failure before output begins, Monsoon advances automatically and records every attempt.

  1. 01

    You set the fallback order

  2. 02

    Automatic failover before output begins

  3. 03

    Every route attempt recorded

Unified access

Integrate once. Route across leading models.

Use one OpenAI-compatible chat-completions interface across the models available to your workspace. Switch models without rebuilding your integration.

POST/v1/chat/completionsView API docs
curl https://api.monsoon.sh/v1/chat/completions \
  -H "Authorization: Bearer $MONSOON_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5.6-terra",
    "models": [
      "anthropic/claude-sonnet-5",
      "google/gemini-3.5-flash"
    ],
    "messages": [{"role": "user", "content": "Your test prompt"}],
    "stream": true
  }'
BASE=https://api.monsoon.sh/v1
AUTH="Authorization: Bearer"
TYPE="Content-Type: application/json"

curl "$BASE/chat/completions" \
  -H "$AUTH $MONSOON_API_KEY" \
  -H "$TYPE" \
  -d '{
    "model":
      "openai/gpt-5.6-terra",
    "models": [
      "anthropic/claude-sonnet-5",
      "google/gemini-3.5-flash"
    ],
    "messages": [{
      "role": "user",
      "content": "Your test prompt"
    }],
    "stream": true
  }'

Auditability

Every routed request leaves a clear trail.

See the model requested, every route attempt, the model served, latency, token usage, estimated upstream cost, and content-logging state—all tied to one request ID.

Requested and served model
Per request
Route attempts in order
Per request
Latency and token usage
Per request
Estimated upstream cost
Per request
Content-logging state
Per request
Request ID
Per request

Vietnam endpoint

A dedicated endpoint for Vietnam.

vn.monsoon.sh already exposes the same OpenAI-compatible contract as the global gateway. The preview does not run inference; Vietnam-resident routing is the next infrastructure milestone.

PreviewVietnam-resident routing is coming soon.
Endpoint
vn.monsoon.sh/v1
Current status
Preview
API contract
OpenAI-compatible

Long-term direction

Gateway first. Frontier capacity for Vietnam.

The gateway brings model access, ordered fallback, and auditability into one control layer—the operating foundation for Vietnam-resident infrastructure and deeper model capability over time.

Sovereign AI means building durable control in Vietnam over deployment, data handling, auditability, infrastructure, and ultimately model capability. The gateway is the first layer of that capacity.