Waytocc Docs

Overview

Get a key

1. 1. Open waytocc.com and register with your email.
2. 2. Go to Wallet and top up any plan.
   (Got an invite code? Enter it at top-up for a first-time bonus.)
3. 3. Go to API Keys and click New key. The key is shown once — copy it immediately.

Replace YOUR_API_KEY in every example below with your key (it looks like sk-...).

Claude chat (Messages API)

Waytocc serves real Claude, fully compatible with Anthropic's POST /v1/messages. Authenticate with x-api-key (or Authorization: Bearer) and set the model to claude-opus-5.

curl https://api.waytocc.com/v1/messages \
  -H "x-api-key: YOUR_CLAUDE_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-opus-5","max_tokens":1024,"stream":true,"messages":[{"role":"user","content":"Introduce yourself in one sentence"}]}'

The response matches Anthropic exactly; text is at content[0].text. Use stream:true — a long non-streamed answer can hit a 524 at the edge after ~100s. Billed per token (input + output); see Usage history.

GPT chat (OpenAI-compatible)

The same gateway serves OpenAI-compatible chat at POST /v1/chat/completions. Set the model to gpt-5.6-sol (gpt-5.6-terra and gpt-5.6-luna are the same generation; gpt-5.4 / gpt-5.4-mini are cheaper) and authenticate with Authorization: Bearer.

curl https://api.waytocc.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_OPENAI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-5.6-sol","stream":true,"messages":[{"role":"user","content":"Introduce yourself in one sentence"}]}'

The response matches OpenAI; text is at choices[0].message.content (delta.content when streaming). Use stream:true. Billed per token.

Gemini chat (native Gemini protocol)

Gemini uses Google's own API shape: POST /v1beta/models/{model}:generateContent, or :streamGenerateContent?alt=sse for streaming. Authenticate with x-goog-api-key and use the bare root as the base URL — the official SDK appends /v1beta itself.

curl "https://api.waytocc.com/v1beta/models/gemini-3-flash:streamGenerateContent?alt=sse" \
  -H "x-goog-api-key: YOUR_GEMINI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"contents":[{"role":"user","parts":[{"text":"Introduce yourself in one sentence"}]}]}'

⚠️ role is required inside contents — omitting it returns 400 INVALID_ARGUMENT rather than anything that names the missing field. The response matches Google exactly; text is at candidates[0].content.parts[0].text. Streaming is SSE with the same frame shape. Prefer streaming. Billed per token.

Using Claude Code

Because the chat API is Anthropic-compatible, the official Claude Code works by pointing its base URL here — two environment variables, using a key from the Claude group.

npm i -g @anthropic-ai/claude-code

export ANTHROPIC_BASE_URL=https://api.waytocc.com
export ANTHROPIC_AUTH_TOKEN=YOUR_CLAUDE_KEY

claude

ANTHROPIC_BASE_URL takes the bare root (no /v1); ANTHROPIC_AUTH_TOKEN takes your key. Other Anthropic-compatible clients (Cline, the SDKs) work the same way.

Using Codex CLI

OpenAI's official Codex CLI connects directly — drop in the two files below (using a key from the ChatGPT group) and run codex in any project.

1. ~/.codex/config.toml (at the top of the file)

model_provider = "OpenAI"
model = "gpt-5.6-sol"
review_model = "gpt-5.6-sol"
model_reasoning_effort = "xhigh"
disable_response_storage = true
network_access = "enabled"
windows_wsl_setup_acknowledged = true
model_context_window = 1000000
model_auto_compact_token_limit = 900000

[model_providers.OpenAI]
name = "OpenAI"
base_url = "https://api.waytocc.com/v1"
wire_api = "responses"
requires_openai_auth = true

2. ~/.codex/auth.json

{ "OPENAI_API_KEY": "YOUR_OPENAI_KEY" }

base_url points at https://api.waytocc.com/v1 with model gpt-5.6-sol. Install with npm i -g @openai/codex. A 403 dispatch means the same thing as in GPT chat: move the key to the ChatGPT group.

Text to image

Give a prompt, get an image. POST /v1/images/generations with a JSON body. Copy any example below, swap in your key, run it.

curl -sS -m 185 https://api.waytocc.com/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "a red panda eating bamboo, studio lighting",
    "n": 1,
    "response_format": "b64_json"
  }' \
  | python3 -c "import sys,json,base64; d=json.load(sys.stdin); open('out.png','wb').write(base64.b64decode(d['data'][0]['b64_json'])); print('saved out.png')"

On success the image is at data[0].b64_json — decode it and write the file.

Image to image

Send an image plus a prompt to edit it. POST /v1/images/edits as multipart/form-data (the image goes in the image field). Note the endpoint is plural: edits.

curl -sS -m 185 https://api.waytocc.com/v1/images/edits \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F model=gpt-image-2 \
  -F [email protected] \
  -F prompt="put the subject in a snowy forest at night" \
  -F response_format=b64_json \
  | python3 -c "import sys,json,base64; d=json.load(sys.stdin); open('out.png','wb').write(base64.b64decode(d['data'][0]['b64_json'])); print('saved out.png')"

Swap input.png for your own file (PNG / JPG / WebP). The response shape matches text-to-image.

Parameters

Response

A successful call returns 200 with the image at data[0].b64_json as raw base64 (no prefix to strip) — decode and write it straight to a file.

{
  "created": 1781184210,
  "data": [
    { "b64_json": "iVBORw0KGgoAAAANSUhEUg..." }
  ],
  "usage": {
    "input_tokens": 8,
    "output_tokens": 0,
    "total_tokens": 8
  }
}

Pricing

Images are billed per image at a flat rate — token usage is not part of the calculation. Prompt length and returned tokens don't change the price.

The tier follows the size of the image you actually get back, not the size parameter you sent: asking for 1024×1024 does not guarantee the 1K tier, and a size that can't be classified falls back to the 2K tier. Text-to-image and image-to-image cost the same; only the actual output size differs.

Chat is billed per token (input + output). Top up at ¥1 = $1; usage and spend are live in Usage history.

Base URL

Each protocol spells its base URL differently, because the official SDKs disagree about who appends the version segment:

The Anthropic and Gemini SDKs append /v1 and /v1beta themselves, so they take the bare root. The OpenAI SDK treats the base URL as already version-qualified, so it needs the /v1.

Common endpoints

Error codes

FAQ

Stuck? Talk to us

Connection problems, errors, balance questions — reach out.