Documentation

Inference APIs gives you access to popular AI models through one HTTPS endpoint and one API key. Chat models use the familiar chat completions request and response format, so existing code needs few changes.

Base URL

https://api.inferenceapis.com/v1

The API follows the OpenAI wire format, so official OpenAI SDKs work by setting this as base_url. /openai/v1 is accepted as well, so Groq-style base URLs work with only the host changed. The model field selects which model handles the request.

Endpoint	Purpose	Models
`POST /v1/chat/completions`	Chat and reasoning (streaming, tools, JSON mode)	Chat models
`POST /v1/audio/speech`	Text to speech → mp3 / wav	Speech models
`POST /v1/audio/transcriptions`	Speech to text (multipart upload)	Transcription models
`GET /v1/models`	List model ids and aliases	—
`GET /v1/voices?model=…`	List voices for a speech model	—

Authentication

Authenticate by sending your secret key in the Authorization header as a Bearer token. Create an account to get a key, then find it on the API Keys page.

Authorization: Bearer $INFERENCE_API_KEY

Treat your key like a password. Don't use it in browser or mobile code, and don't commit it to source control. Load it from an environment variable or a secrets manager.

Make a request

Send a list of messages and receive the model's reply. This example uses Llama 3.3 70B Instruct:

curl https://api.inferenceapis.com/v1/chat/completions \
  -H "Authorization: Bearer $INFERENCE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Hello! What can you do?"}
    ]
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.inferenceapis.com/v1",
    api_key=os.environ["INFERENCE_API_KEY"],
)
response = client.chat.completions.create(
    model="meta-llama/Llama-3.3-70B-Instruct-Turbo",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello! What can you do?"},
    ],
)
print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.inferenceapis.com/v1",
  apiKey: process.env.INFERENCE_API_KEY,
});
const response = await client.chat.completions.create({
  model: "meta-llama/Llama-3.3-70B-Instruct-Turbo",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "Hello! What can you do?" },
  ],
});
console.log(response.choices[0].message.content);

The reply text is in choices[0].message.content, and usage reports prompt and completion token counts. Each model page lists all supported parameters.

Models

Model	Model ID	Type	Status
GPT-OSS 120B	`openai/gpt-oss-120b`	Chat	Available
DeepSeek V4 Flash	`deepseek-ai/DeepSeek-V4-Flash`	Chat	Available
DeepSeek V4.1 Flash	`deepseek-ai/DeepSeek-V4.1-Flash`	Chat	Available
GLM 5.3 Flash	`zai-org/GLM-5.3-Flash`	Chat	Available
Llama 3.3 70B Instruct	`meta-llama/Llama-3.3-70B-Instruct-Turbo`	Chat	Available
GPT-OSS 20B	`openai/gpt-oss-20b`	Chat	Coming soon
Qwen3-VL 8B	`Qwen/Qwen3-VL-8B-Instruct`	Chat	Coming soon
Kokoro 82M	`hexgrad/Kokoro-82M`	Speech	Available
Orpheus 3B	`canopylabs/orpheus-3b-0.1-ft`	Speech	Available
Whisper Large v3	`openai/whisper-large-v3`	Transcription	Available
Parakeet TDT 0.6B v3	`nvidia/parakeet-tdt-0.6b-v3`	Transcription	Available
OmniParser V2	`omniparser2`	Vision	Unavailable

Streaming

Set "stream": true on a chat request to receive the response as server-sent events. Each event is a data: line containing a JSON chunk with the next piece of text in choices[0].delta.content.

Errors

When a request can't be completed, the response body contains an error object with a human-readable message. Always check for it before reading choices.

Error response

{
  "error": {
    "message": "Invalid JSON payload",
    "type": "invalid_request_error"
  }
}

Cause	What to do
Invalid JSON	Check the request body is valid JSON and `Content-Type` is `application/json`.
Unknown model	Use a model ID from the models list.
Service unavailable	The model is temporarily unavailable. Retry with exponential backoff or switch models.
Insufficient credits	Add credits to your account.

Billing

Usage is paid from prepaid credits. New accounts start with free credit, and you can top up at any time from the Billing page.

API reference