Create an API key
Open API Keys, create a key, and copy it immediately. The full secret is shown only once.
Go to API keysKnowledge Base
Everything needed to connect chat, API keys, tools, pricing, and limits.
2Choose your tool
Cursor
OpenAI-compatible provider
- Open Cursor Settings -> ModelsAdd a new OpenAI Compatible provider.
- Paste connection data
OpenAI API Key: YOUR_API_KEYOpenAI Base URL: https://apiclaude.org/v1 - Set model aliases
claude-opus-4-7claude-sonnet-4-6claude-haiku-4-5
Claude Code
Official CLI
- Install Claude Code
npm install -g @anthropic-ai/claude-code - Set environment variables
export ANTHROPIC_BASE_URL="https://apiclaude.org/v1"export ANTHROPIC_API_KEY="YOUR_API_KEY" - Run it in your project
claude
Kilo Code
VS Code extension
Some IDE agents do not forward images to custom providers. Text, code, and normal chat work correctly.
- Install the extensionOpen VS Code Marketplace and find Kilo Code.
- Paste provider settings
Anthropic Base URL: https://apiclaude.org/v1API Key: YOUR_API_KEY - Choose a model
claude-sonnet-4-6
Cline / Roo
VS Code agent
Images may be limited by the client itself. For text and code use OpenAI Compatible mode.
- Choose provider
OpenAI Compatible - Paste Base URL and API key
https://apiclaude.org/v1YOUR_API_KEY - Set Model ID
claude-haiku-4-5
Continue
VS Code / JetBrains
name: ApiClaude
version: 1.0.0
schema: v1
models:
- name: Claude Sonnet 4.6
provider: openai
model: claude-sonnet-4-6
apiBase: https://apiclaude.org/v1
apiKey: YOUR_API_KEY
- name: Claude Haiku 4.5
provider: openai
model: claude-haiku-4-5
apiBase: https://apiclaude.org/v1
apiKey: YOUR_API_KEYOpenCode
Terminal agent
- Install OpenCode
npm install -g opencode-ai - Create or edit config
"baseURL": "https://apiclaude.org/v1""apiKey": "YOUR_API_KEY" - Run it
opencode
Models
Manual model choice always uses the selected model. Auto mode is the separate economy router.
AutoDefault
cligate-chat
Routes simple requests to cheaper models and keeps explicit model selection untouched.Claude Opus 4.7NEW
claude-opus-4-7
Claude Opus 4.6
claude-opus-4-6
Claude Opus 4.7 1M1M
claude-opus-4-7-1m
Gemini 3.1 Pro PreviewNEW
gemini-3-1-pro-preview
Claude Sonnet 4.6 1M1M
claude-sonnet-4-6-1m
GPT 5.5NEW
gpt-5-5
GPT 5.4ДЁШЕВО
gpt-5-4
Claude Opus 4.6 1M1M
claude-opus-4-6-1m
Claude Opus 4.5
claude-opus-4-5
Claude Sonnet 4.6
claude-sonnet-4-6
Claude Sonnet 4.5
claude-sonnet-4-5
Claude Haiku 4.5
claude-haiku-4-5
Model Pricing
Cost per 1M tokens. Billing is charged in RUB using the current service rate 72.3497 · CBR.
| Model | Input | Output | Approx. RUB input | Approx. RUB output |
|---|---|---|---|---|
Claude Opus 4.7 NEWclaude-opus-4-7 |
50¢ | 50¢ | 36,17 ₽ | 36,17 ₽ |
Claude Opus 4.6claude-opus-4-6 |
50¢ | 50¢ | 36,17 ₽ | 36,17 ₽ |
Claude Opus 4.7 1M 1Mclaude-opus-4-7-1m |
150¢ | 150¢ | 108,52 ₽ | 108,52 ₽ |
Gemini 3.1 Pro Preview NEWgemini-3-1-pro-preview |
50¢ | 50¢ | 36,17 ₽ | 36,17 ₽ |
Claude Sonnet 4.6 1M 1Mclaude-sonnet-4-6-1m |
120¢ | 120¢ | 86,82 ₽ | 86,82 ₽ |
GPT 5.5 NEWgpt-5-5 |
50¢ | 50¢ | 36,17 ₽ | 36,17 ₽ |
GPT 5.4 ДЁШЕВОgpt-5-4 |
25¢ | 25¢ | 18,09 ₽ | 18,09 ₽ |
Claude Opus 4.6 1M 1Mclaude-opus-4-6-1m |
150¢ | 150¢ | 108,52 ₽ | 108,52 ₽ |
Claude Opus 4.5claude-opus-4-5 |
50¢ | 50¢ | 36,17 ₽ | 36,17 ₽ |
Claude Sonnet 4.6claude-sonnet-4-6 |
40¢ | 40¢ | 28,94 ₽ | 28,94 ₽ |
Claude Sonnet 4.5claude-sonnet-4-5 |
40¢ | 40¢ | 28,94 ₽ | 28,94 ₽ |
Claude Haiku 4.5claude-haiku-4-5 |
40¢ | 40¢ | 28,94 ₽ | 28,94 ₽ |
API Reference
All private endpoints require Authorization: Bearer YOUR_API_KEY.
https://apiclaude.org/v1
/chat/completions
OpenAI-compatible Chat Completions API. Use it with OpenAI SDKs, Cursor, Continue, Cline, and similar clients.
auth/messages
Anthropic-style Messages API through the proxy when the client sends Anthropic payloads.
auth/models
Available models in OpenAI format.
auth/balance
Current balance for the account attached to the API key.
auth/messages/count_tokens
Token counting without a model call when supported by the upstream provider.
authFAQ
Frequently asked questions
Why can one request use many tokens?
IDE agents often send chat history and project files with every request. Use shorter sessions, summaries, and context compaction.
Which Claude version is used?
The account shows public model aliases. If you select a specific model, ApiClaude does not substitute it. Auto mode is separate.
Do you inflate token counts?
No. Charges are based on provider and LiteLLM usage fields saved in logs.
How do I create an API key?
Open API Keys, click Create key, set name and optional limit. The full key is shown only once.
How do I top up balance?
Click Top up, choose amount and method. Before automatic payments are connected, requests are confirmed by an admin.
Why is 1M context more expensive?
Large-context models use more expensive account capacity and should be selected only when a task really needs huge files or long history.
Useful Tips
How to keep quality high and spend under control.
Compress history
Long context is expensive. Start a new chat for new tasks and keep only what matters.
/compactUse separate keys
Separate keys make limits and spending much easier to track.
Track logs
Logs show request source, model, token count, cost, latency, and status.
Pick the right model
Use Haiku or Sonnet for everyday work. Keep Opus and 1M context for hard tasks.
Set key limits
A RUB spend limit protects the balance from unexpected client behavior.
Ask support
If a model is unavailable or the client fails, support can see request status and help faster.