Early Access · Architecture preview

Free AI models.
One API.

Access free LLMs through a simple OpenAI-compatible API. No credit card. No model hunting. Just build.

Early Access

The API is being wired up.

TokenForFree is not issuing API keys yet. We are building the routing and model-supply layer now—without pretending the endpoint is live.

WebsiteLive
Model feedIn progress
API keysComing next

Email collection will be added with the API backend. No address is collected on this static release.

Explore Models
$0to start
1API
OpenAIcompatible
QUICKSTART / OPENAI COMPATIBLEILLUSTRATIVE
curl https://api.tokenforfree.com/v1/chat/completions \
  -H "Authorization: Bearer $TFF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tff/free",
    "messages": [
      { "role": "user", "content": "Hello!" }
    ]
  }'
Illustrative request · API not live yet
FREE MODELS AVAILABLE NOWDEMO DATA
Autotff/free
Fast · Reasoning Early AccessFREE
Fast Pooltff/free-fast
Fast Early AccessFREE
Reasoning Pooltff/free-reasoning
Reasoning Early AccessFREE
Coding Pooltff/free-coding
Coding · Tools Early AccessFREE
View all free models
02 / THE CONTRACT

Less ceremony around inference.

One interface between your application and fragmented free supply.

01

Always free options

The router is designed to keep a free path available when eligible supply exists.

02

One API

Switch routing profiles without rewriting your application.

03

OpenAI compatible

Keep the SDKs and request shapes you already know.

04

Automatic routing

Let TokenForFree find an eligible free model for each request.

05

No credit card

Start without a billing setup or a credit balance.

06

Built for developers

API first, inspectable behavior, and no unnecessary complexity.

03 / ROUTING

One request. A free route.

01Your App
02TokenForFree API
03Free Router
04Available Free Model

TokenForFree is designed to check available free inference sources and route requests to an eligible model.

tff/freetff/free-fasttff/free-reasoning
Routing profiles shown here are an architecture preview. The backend is not live yet.
04 / QUICKSTART

Keep your existing SDK.

Change the base URL and model ID. The examples are ready for backend integration.

Read the docs
from openai import OpenAI

client = OpenAI(
    base_url="https://api.tokenforfree.com/v1",
    api_key="YOUR_TFF_API_KEY",
)

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

print(response.choices[0].message.content)
Illustrative request · API not live yet
05 / SETUP

Less setup. More building.

Traditional API setup

  • Choose provider
  • Create account
  • Add card
  • Get API key
  • Track credits
  • Switch SDK

TokenForFree

  • One API
  • No card
  • Free models
  • Automatic routing
  • OpenAI compatible
EARLY ACCESS

Start building for free.

No credit card required. API access is coming next.

Early Access

The API is being wired up.

TokenForFree is not issuing API keys yet. We are building the routing and model-supply layer now—without pretending the endpoint is live.

WebsiteLive
Model feedIn progress
API keysComing next

Email collection will be added with the API backend. No address is collected on this static release.