[!IMPORTANT]
I'm working to keep the API online at all times.
Join our community to be among the first to know
when the project is fully up and running.

💬 Community: https://telegram.me/openapis

OpenAPIs

Free Claude API + free OpenAI API in one unified proxy. Drop-in compatible with the official Anthropic SDK, OpenAI SDK, Claude Code CLI, OpenCode, and any OpenAI-compatible client. 100+ models — Claude Opus 4.7, Sonnet 4.6, Haiku 4.5, GPT-5.5, GPT-5.4-mini, GPT-5.3-codex — totally free during the open beta.

OpenAPIs is a unified proxy to OpenAI and Anthropic. Drop-in compatible with their official SDKs — just swap the base_url, keep your code.

🎁 Open Beta — Free

Currently in open beta. Completely free access while the beta lasts.

Why OpenAPIs

	Your existing code with the official SDKs keeps working — zero refactor
	End-to-end SSE. Tokens arrive instantly, no buffering
	Forget juggling multiple accounts, dashboards, and bills
	OpenAI + Anthropic + Claude Code under one URL

Set up your env vars (recommended)

The official OpenAI and Anthropic SDKs read these variables automatically, so you never have to pass base_url or api_key in code.

# OpenAI
export OPENAI_API_KEY=admin
export OPENAI_BASE_URL=https://api.openapis.online/openai

# Anthropic
export ANTHROPIC_API_KEY=admin
export ANTHROPIC_BASE_URL=https://api.openapis.online/anthropic

# Windows PowerShell
$env:OPENAI_API_KEY = "admin"
$env:OPENAI_BASE_URL = "https://api.openapis.online/openai"
$env:ANTHROPIC_API_KEY = "admin"
$env:ANTHROPIC_BASE_URL = "https://api.openapis.online/anthropic"

Quickstart

OpenAI

curl — with streaming

curl https://api.openapis.online/openai/v1/chat/completions \
  -H "Authorization: Bearer admin" \
  -H "content-type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "max_completion_tokens": 1024,
    "stream": true,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Python — official SDK

from openai import OpenAI

client = OpenAI(
    api_key="admin",
    base_url="https://api.openapis.online/openai/v1",
)

# Without streaming
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

# With streaming
with client.chat.completions.stream(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello"}],
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

Anthropic

curl — with streaming

curl -N https://api.openapis.online/anthropic/v1/messages \
  -H "x-api-key: admin" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-opus-4-7",
    "max_tokens": 1024,
    "stream": true,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Python — official SDK

from anthropic import Anthropic

client = Anthropic(
    api_key="admin",
    base_url="https://api.openapis.online/anthropic",
)

# Without streaming
response = client.messages.create(
    model="claude-opus-4-7",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.content[0].text)

# With streaming
with client.messages.stream(
    model="claude-opus-4-7",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

Claude Code (CLI)

# Linux / macOS / WSL
export ANTHROPIC_BASE_URL="https://api.openapis.online/anthropic"
export ANTHROPIC_API_KEY="admin"

claude

# Windows PowerShell
$env:ANTHROPIC_BASE_URL = "https://api.openapis.online/anthropic"
$env:ANTHROPIC_API_KEY = "admin"

claude

OpenCode (CLI)

# Linux / macOS / WSL
export ANTHROPIC_BASE_URL="https://api.openapis.online/anthropic"
export ANTHROPIC_API_KEY="admin"
export OPENAI_BASE_URL="https://api.openapis.online/openai"
export OPENAI_API_KEY="admin"

opencode

# Windows PowerShell
$env:ANTHROPIC_BASE_URL = "https://api.openapis.online/anthropic"
$env:ANTHROPIC_API_KEY = "admin"
$env:OPENAI_BASE_URL = "https://api.openapis.online/openai"
$env:OPENAI_API_KEY = "admin"

opencode

Available models (100+)

OpenAI — endpoint /openai

• gpt-5.5 · gpt-5.5-pro
• gpt-5.4-mini · gpt-5.4-nano · gpt-5.4-pro · gpt-5.4-thinking
• gpt-5.3-codex · gpt-5.3-instant
• … the entire GPT-5 family with all snapshots

Anthropic — endpoint /anthropic

• claude-opus-4-7 · claude-opus-4-6
• claude-sonnet-4-6 · claude-sonnet-4-5
• claude-haiku-4-5 · claude-haiku-4
• … the full Claude catalog (3.x → 4.x)

Compatible stack

Anything that speaks the OpenAI or Anthropic Messages API format works out of the box:

• OpenAI SDK (Python, Node, Go, Java, .NET)
• Anthropic SDK (Python, Node)
• Claude Code — Anthropic's official CLI
• LangChain / LangGraph
• Vercel AI SDK
• LiteLLM
• OpenRouter-style routers
• Continue.dev, Cursor, Cline, and any other IDE with a configurable base URL

FAQ

How do I use Claude Code for free?

Set ANTHROPIC_BASE_URL and ANTHROPIC_API_KEY to point at OpenAPIs (see Claude Code (CLI)). While the open beta lasts, every Claude Code request goes through the proxy at no cost — no Anthropic billing key needed on your side.

Is the Anthropic API really free here?

Yes. During the open beta, the proxy gives you free Anthropic API access to the entire Claude catalog: claude-opus-4-7, claude-sonnet-4-6, claude-haiku-4-5, plus the rest of 3.x → 4.x. You never pass an Anthropic key — only the beta key.

Can I use this as a free OpenAI API?

Yes. The same proxy serves OpenAI-compatible chat completions for the full GPT-5 family. Point your OpenAI SDK at /openai/v1 and it behaves identically to the official endpoint — client.chat.completions.create(...) works unchanged.

Does Claude Code work without an Anthropic API key?

Through OpenAPIs, yes. Claude Code accepts any Anthropic-compatible endpoint, so by setting ANTHROPIC_BASE_URL to OpenAPIs you skip the official key entirely. This is the same trick proxies like claude-code-proxy, anthropic-proxy, and claude-code-router use, but hosted for you.

Is OpenCode supported?

Yes. OpenCode reads the standard OPENAI_API_KEY, OPENAI_BASE_URL, ANTHROPIC_API_KEY, and ANTHROPIC_BASE_URL env vars — set all four and run opencode. Full setup in OpenCode (CLI).

Which models can I call?

Over 100, covering OpenAI text/reasoning models (GPT-5.5, GPT-5.5-pro, GPT-5.4-mini, GPT-5.4-nano, GPT-5.4-thinking, GPT-5.3-codex, plus dated snapshots) and the full Claude 3.x → 4.x catalog. Image, video, and audio models are intentionally excluded. See Available models.

How is this different from OpenRouter, LiteLLM, or claude-code-proxy?

OpenRouter aggregates many providers behind a unified billing page; LiteLLM is a Python library you self-host; claude-code-proxy is a translation layer you run locally. OpenAPIs is a hosted proxy that keeps the official OpenAI and Anthropic API shapes intact, so any client built for those APIs (including Claude Code, which is strict about the Anthropic format) works without translation layers or local processes.

What about rate limits?

None during the open beta. Be considerate — the beta key is shared.

Does it support streaming?

Yes, native SSE end-to-end on both /openai and /anthropic endpoints. See the curl -N and Python client.*.stream() examples above.

Contact

• 🐙 GitHub Issues — bugs, feature requests, and beta access

Keywords

Free Claude API · Free Anthropic API · Free OpenAI API · Free Claude Code · Claude Code free · Claude Code without API key · Claude Code proxy · Anthropic proxy · Anthropic to OpenAI · OpenAI compatible proxy · OpenAI reverse proxy · OpenCode proxy · OpenCode free · GPT-5 free · GPT-5.5 free · Claude Opus free · Claude Sonnet free · Claude Haiku free · LLM proxy · AI gateway · LiteLLM alternative · OpenRouter alternative · Free LLM API · Unified AI API · OpenAI Anthropic gateway

License

MIT — the landing page and examples are open source. The service behind the proxy is proprietary.