Skip to content

One API.Every model.

Frontier and open weight models through a single gateway.

Get API key

Frontier and open-weight models, through a single gateway.

  • OpenAI
  • GPT-4o
  • Anthropic
  • Claude Sonnet 4
  • Open-weight
  • GPT OSS 120B

Start in minutes.Then build.

  1. 1. Choose a model

    From the catalogue, with prices

    Pick an enabled model and see its price, limits and features before you send anything.

  2. 2. Create a key

    Scoped to your application

    Allow specific models, set a budget and expiry, then rotate or revoke the key from the dashboard.

  3. 3. Make the call

    With your existing client

    Point your existing SDK or HTTP client at SingleAPI and send a chat completion.

Change the model.Keep the interface.

Use the same chat-completion endpoint across supported providers. Pick a model below to see the request.

Check availability
Python · OpenAI SDK 
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["SINGLEAPI_API_KEY"],
    base_url="https://api.singleapi.ai/v1",
)

stream = client.chat.completions.create(
    model="openai/gpt-4o-mini",
    messages=[{"role": "user", "content": "Say hello."}],
    stream=True,
)

for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)
Use cURL instead
cURL
curl "https://api.singleapi.ai/v1/chat/completions" \
  -H "Authorization: Bearer $SINGLEAPI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4o-mini",
    "messages": [{"role": "user", "content": "Say hello."}],
    "stream": true
  }' --no-buffer

Example only. Select an enabled model and allow it on your API key before sending a request. Streaming depends on model support.

Choose howyou connect.

Use managed access to pay through SingleAPI, or connect a provider account you already have.

Explore BYOK

Managed credits

Add prepaid USD credits. SingleAPI handles provider authentication and settles your wallet charge against recorded usage.

Usage billed to
Your SingleAPI wallet
Request credentials
Your SingleAPI API key

Review current customer rates in the model catalogue.

Your provider key

Store an encrypted provider credential and select it on a request. Your provider bills you directly. SingleAPI does not debit your wallet.

Usage billed to
Your provider account
Request credentials
SingleAPI key plus stored credential ID

The model must match your credential's provider.

Before youmake the call.

Will my existing SDK work?

Use an OpenAI-compatible SDK or HTTP client with SingleAPI's base URL. Compatibility covers our documented endpoints and request fields, rather than every feature of the OpenAI API.

Check the API reference
Which models can I use?

The model catalogue shows enabled models, customer prices, limits and features. Your API key must also allow the model. Model availability can change.

See models and pricing
Is streaming supported?

Yes, for models with streaming enabled in the catalogue. Set stream to true to receive incremental output. Function tools and JSON formats also depend on the selected model's enabled features.

Read about streaming
Do you store my prompts?

Content logging is off by default. When you opt in on a key, future prompts and responses are encrypted and expire after 30 days. Provider processing and backup retention are separate.

Understand logging controls
How does billing work?

Managed requests reserve funds before execution and settle using recorded usage and the applied customer price. BYOK requests do not debit your SingleAPI wallet.

Read about billing

One gateway.Every control.

  1. Any model, one interfaceThe same request everywhere.

    Call every supported model through one endpoint, with the same request shape every time.

  2. Open-weight models, no platform feeThe platform fee is for closed-source only.

    Call hosted open-source models with no platform fee at all. The 4% platform fee applies only to closed-source models.

  3. Managed or your own keyPay the way you prefer.

    Pay through SingleAPI with prepaid credits, or bring your own provider key and let your provider bill you.

  4. Budgets and scopes per keyLimits you set and change.

    Allow specific models, set a budget and an expiry on each key, then rotate or revoke it anytime.

  5. Usage and cost recordsSeparate from prompt content.

    See tokens, charges and timing by key and model, kept apart from optional prompt content.

  6. Streaming, tools and JSONOn the models that support them.

    Stream responses and use function tools or JSON output on models with those features enabled.

  7. Opt-in prompt loggingEncrypted and time-limited.

    Turn on encrypted content logs per key when you need them. Captured content expires after 30 days.

Create your key.Make the first call today.

Choose an enabled model, scope a key to your application and connect your existing client.

Get API key