Quick Start

Try It Online

Want to try TierFlow without installing anything? Send a request with cURL:

curl https://api.tierflow.dev/v1/chat/completions \
  -H "Authorization: Bearer your-tierflow-key" \
  -H "Content-Type: application/json" \
  -d '{"model": "auto", "messages": [{"role": "user", "content": "Hello"}]}'

Integrate TierFlow

Prerequisites

• Familiarity with the command line
• Python 3.8+ or Node.js 18+
• A TierFlow API Key

Get Your API Key

Go to the TierFlow Console and create a new key on the "API Keys" page:

TIERFLOW_API_KEY=tf-sk-xxxxxxxxxxxxxxxxxxxxxxxx

Install SDK

Python

pip install openai

Node.js

npm install openai

TierFlow is compatible with the OpenAI SDK — no extra dependencies needed.

Replace Base URL

Just change base_url and api_key. No other code changes required.

Python

from openai import OpenAI

client = OpenAI(
    api_key="tf-sk-xxxxxxxxxxxxxxxxxxxxxxxx",
    base_url="https://api.tierflow.dev/v1"
)

response = client.chat.completions.create(
    model="auto",
    messages=[{"role": "user", "content": "Explain vector databases in one sentence"}]
)

print(response.choices[0].message.content)

Node.js

import OpenAI from "openai";

const client = new OpenAI({
    apiKey: "tf-sk-xxxxxxxxxxxxxxxxxxxxxxxx",
    baseURL: "https://api.tierflow.dev/v1"
});

const response = await client.chat.completions.create({
    model: "auto",
    messages: [{ role: "user", content: "Explain vector databases in one sentence" }]
});

console.log(response.choices[0].message.content);

cURL

curl https://api.tierflow.dev/v1/chat/completions \
  -H "Authorization: Bearer tf-sk-xxxxxxxxxxxxxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto",
    "messages": [{"role": "user", "content": "Explain vector databases in one sentence"}]
  }'

That's it. Your requests are now intelligently routed through TierFlow.

Routing Modes

The model field supports multiple routing strategies. Start with auto:

Mode	model value	Description
Auto Route	auto	Automatically selects the optimal model based on task complexity (recommended)

For more routing strategies (cost-first, quality-first, specific model, etc.), see Routing Strategy.

response = client.chat.completions.create(
    model="auto",
    messages=[{"role": "user", "content": "Hello"}]
)

View Routing Results

Each response includes routing metadata showing which model was actually used:

{
  "model": "gpt-4o-mini",
  "tierflow": {
    "requested_model": "auto",
    "routed_model": "gpt-4o-mini",
    "route_reason": "simple_query",
    "estimated_cost": "$0.000042",
    "saved_vs_flagship": "82%"
  }
}

Streaming

Works exactly like the OpenAI API — just use stream=True:

stream = client.chat.completions.create(
    model="auto",
    messages=[{"role": "user", "content": "Write a short essay about AI"}],
    stream=True
)

for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

FAQ

Does it affect response quality?

No. The routing engine evaluates each request's complexity and only downgrades when a lighter model can handle it. You can set quality thresholds in the console.

Which models are supported?

OpenAI, Anthropic, Google, DeepSeek, Alibaba Cloud, and more. See Model List for the full list.

How do I set budget limits?

Set daily/monthly limits in the console under "Budget Management". The API returns 429 when limits are exceeded.

Next Steps

If you haven't read the Introduction, we recommend starting there.

• Routing Strategy — Deep dive into how the routing engine works.
• Model List — All supported models and pricing.
• API Reference — Complete API documentation.