Drop-in OpenAI-style API · local & cloud

Trinix API

Works like OpenAI with the clients you already use—often cheaper—and one surface for local and cloud models. Production endpoints for chat, generation, embeddings, and web search, with limits that behave in the real world. New here? Use the browser playground to try free models before you create an account.

Base URL api.trinix.gg Rate limit 100/min per IP & per key/session Body 1 MB max

🧠 Why this matters

Fresh, grounded answers shouldn’t live behind a separate product lane or an enterprise upsell. If you expose search the same way you expose everything else—a plain HTTP call—

POST /api/v1/web-search

👉 That’s just a basic feature. Same X-API-Key as /api/v1/chat, same integration mindset, no second vendor or special SDK for “grounding.” Builders wire it in like any other endpoint and move on.

Authentication

Use X-API-Key on native endpoints and either X-API-Key or bearer auth on OpenAI-compatible endpoints.

X-API-Key: YOUR_API_KEY
Authorization: Bearer YOUR_API_KEY

Endpoints

Native API

GET /api/v1/models
GET /api/v1/queue-status
POST /api/v1/chat
POST /api/v1/generate
POST /api/v1/embed
POST /api/v1/web-search

OpenAI-compatible

POST /v1/chat/completions
GET /v1/models

Task API

POST /task/v1/chat/completions
GET /task/v1/models

Models

⚡ trinix-chat free
💻 trinix-code free
🧠 trinix-reason free
📐 trinix-embed free
🚀 trinix-pro paid
💻 trinix-coder-pro paid
🧠 trinix-reason-pro paid
🚀 trinix-ultra paid

Quick start

Native chat

curl https://api.trinix.gg/api/v1/chat \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"trinix-chat","messages":[{"role":"user","content":"hello"}]}'

OpenAI-compatible chat

curl https://api.trinix.gg/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"trinix-chat","messages":[{"role":"user","content":"hello"}]}'

Python SDK

from openai import OpenAI

client = OpenAI(base_url="https://api.trinix.gg/v1", api_key="YOUR_API_KEY")
resp = client.chat.completions.create(
    model="trinix-chat",
    messages=[{"role": "user", "content": "hello"}]
)
print(resp.choices[0].message.content)

Streaming

Native: NDJSON chunks ({"delta":"..."}, then {"done":true}).
OpenAI-compatible: SSE frames ending with data: [DONE].

Limits

Prompt/input cap by model and plan (up to 200k chars for enterprise paid flow).
Request body maximum: 1 MB.
Rate limits: 100 requests/minute per client IP and again per API key (or per browser session when no key). Both apply.
Paid models enforce daily and weekly quotas.
Web search (POST /api/v1/web-search) daily quota by plan:

Plan	Web search
Free	❌ Disabled
Pro	10/day
Enterprise	50/day default · typically 50–100/day on contract

Queue full/timeout returns 429.