Connect IDEs
Use one ApiClaude key in Cursor, Claude Code, OpenCode, Continue, Cline/Roo, Codex, and SDKs.
ApiClaude
ApiClaude Docs
Connect IDEs, web chat, voice input, agents, billing, and API tools from one clean guide.
Web chat
Ask from the browser with saved history, model selection, voice input, and balance tracking.
Open chatVoice input
Dictate messages in web chat or Hermes. ApiClaude transcribes audio and sends it as a normal prompt.
Try voiceAI assistants
Launch a personal assistant for research, drafts, monitoring, and repeatable work from the same account.
Open agents
IDE setup
Your ApiClaude connection values
Copy these into your IDE, local agent, or SDK client.
https://apiclaude.org/v1
YOUR_API_KEY
claude-sonnet-5
claude-opus-4-8
claude-fable-5
claude-haiku-4-5
https://apiclaude.org/v1/models
Create an API key
The full secret is shown once. Use a separate key for every IDE, bot, or agent so limits and logs stay clean.
Choose model consciously
Sonnet 4.6 is the default. Use Haiku for cheaper everyday work and Opus for tasks that really need stronger reasoning.
Use voice when it is faster
Voice input is available in the web chat and agent chat, then the recognized text is sent through the same model and billing logic.
IDE and agent clients
Connect ApiClaude to your editor or terminal client with the compatible API mode that client supports.
Paid models support text, code, and image input in Chat, assistants, and compatible API clients when the client sends image_url. Free models support text and code files only.
Cursor
- Open model settingsAdd ApiClaude as a compatible API connection.
- Paste connection data
OpenAI API Key: YOUR_API_KEYOpenAI Base URL: https://apiclaude.org/v1 - Add model aliases
claude-fable-5claude-sonnet-5claude-opus-4-8claude-opus-4-7claude-sonnet-4-6claude-sonnet-4-5claude-opus-4-6claude-opus-4-5claude-haiku-4-5gpt-5-5gpt-5-4
ApiClaude Desktop
- Download the app
https://apiclaude.org/downloads/ApiClaudeDesktopBridge.exe - Paste your ApiClaude keyOpen CLI Router, click Fetch models, then Start router.
- Use the local Base URL
Base URL: http://127.0.0.1:18441/v1API Key: any non-empty valueModel: choose from Fetch models - Client download linksCodex · Claude Code · Gemini CLI · Hermes
Claude Code
- Install Claude Code
irm https://claude.ai/install.ps1 | iexnpm install -g @anthropic-ai/claude-codeOfficial install guide - Set environment variables
export ANTHROPIC_BASE_URL="https://apiclaude.org/v1"export ANTHROPIC_API_KEY="YOUR_API_KEY" - Run it in your project
claude
Codex
- Install Codex
powershell -ExecutionPolicy ByPass -c "irm https://chatgpt.com/codex/install.ps1 | iex"Official install guide - Choose a custom connectionUse OpenAI-compatible mode in the desktop app or terminal config.
- Paste connection fields
Base URL: https://apiclaude.org/v1API Key: YOUR_API_KEYModel: claude-sonnet-5 - Terminal config
model = "claude-sonnet-5" model_provider = "apiclaude" [model_providers.apiclaude] name = "ApiClaude" base_url = "https://apiclaude.org/v1" env_key = "APICLAUDE_API_KEY" wire_api = "responses"
Gemini CLI
- Install Gemini CLI
npm install -g @google/gemini-cliOfficial install guide - ApiClaude local endpoint
Base URL: http://127.0.0.1:18441/v1API Key: local-routerModel: choose from Fetch models - Use when supportedGemini CLI uses Google auth by default. ApiClaude Desktop keeps a local endpoint ready for compatible Gemini settings.
Hermes
- Install Hermes
iex (irm https://hermes-agent.nousresearch.com/install.ps1)Official install guide - Set custom endpoint
OPENAI_BASE_URL=http://127.0.0.1:18441/v1OPENAI_API_KEY=local-router - Config model
model: provider: custom model: claude-sonnet-5 base_url: http://127.0.0.1:18441/v1 api_key: local-router
OpenCode
- Install OpenCode
curl -fsSL https://opencode.ai/install | bashnpm install -g opencode-ai - Create or edit config
"baseURL": "https://apiclaude.org/v1""apiKey": "YOUR_API_KEY""model": "claude-sonnet-5" - Run it
opencode
VS Code native BYOK
Use this for the built-in VS Code Chat custom model flow. Save the file, reload VS Code, then pick ApiClaude Opus 4.8 in Agent mode.
- Open the config file
%APPDATA%\Code\User\chatLanguageModels.json - Paste this JSON
[
{
"name": "ApiClaude",
"vendor": "customendpoint",
"apiKey": "YOUR_API_KEY",
"apiType": "chat-completions",
"models": [
{
"id": "claude-fable-5",
"name": "Claude Fable 5",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-sonnet-5",
"name": "Claude Sonnet 5",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-opus-4-8",
"name": "Claude Opus 4.8",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-opus-4-7",
"name": "Claude Opus 4.7",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-sonnet-4-6",
"name": "Claude Sonnet 4.6",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-sonnet-4-5",
"name": "Claude Sonnet 4.5",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-opus-4-6",
"name": "Claude Opus 4.6",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-opus-4-5",
"name": "Claude Opus 4.5",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "claude-haiku-4-5",
"name": "Claude Haiku 4.5",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 184000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "gpt-5-5",
"name": "ChatGPT 5.5",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 128000,
"maxOutputTokens": 8000,
"streaming": true
},
{
"id": "gpt-5-4",
"name": "GPT 5.4",
"url": "https://apiclaude.org/v1/chat/completions",
"toolCalling": true,
"vision": true,
"maxInputTokens": 128000,
"maxOutputTokens": 8000,
"streaming": true
}
]
}
]- Reload VS Code
Developer: Reload Window - If an old key was used beforeOpen Chat model picker, remove the old Custom Endpoint entry, then reload and add the current ApiClaude key again.
Continue
name: ApiClaude
version: 1.0.0
schema: v1
models:
- name: Claude Sonnet 5
provider: openai
model: claude-sonnet-5
apiBase: https://apiclaude.org/v1
apiKey: YOUR_API_KEY
capabilities:
- tool_use
roles:
- chat
- edit
- apply
- summarize
defaultCompletionOptions:
temperature: 0
maxTokens: 1600Cline / Roo
- Provider
OpenAI Compatible - Base URL and API key
https://apiclaude.org/v1YOUR_API_KEY - Model ID
claude-sonnet-5
Kilo Code
Some IDE assistants limit image input in compatible API connections. Text, code, and normal chat work through ApiClaude.
- Install extensionVS Code Marketplace · Install guide
- Connection settings
Provider: OpenAI CompatibleBase URL: https://apiclaude.org/v1API Key: YOUR_API_KEYModel ID: claude-sonnet-5
More clients
Use the same fields in any compatible client: connection type, Base URL, API key, and model id.
- Aider
OPENAI_API_BASE=https://apiclaude.org/v1OPENAI_API_KEY=YOUR_API_KEYaider --model openai/claude-sonnet-5 - n8n
OpenAI Chat Model · Base URL: https://apiclaude.org/v1API Key: YOUR_API_KEYModel: claude-sonnet-5 - LangChain / LlamaIndex / PydanticAI
base_url="https://apiclaude.org/v1"api_key="YOUR_API_KEY"model="claude-sonnet-5" - Open WebUI / LibreChat
OPENAI_API_BASE_URL=https://apiclaude.org/v1OPENAI_API_KEY=YOUR_API_KEY - Vercel AI SDK / TanStack AI
baseURL: "https://apiclaude.org/v1"apiKey: process.env.APICLAUDE_API_KEY
Managed assistant
Open agents
Hermes lives inside the same ApiClaude account
Balance, keys, models, limits, and usage history stay shared. Assistant memory, modules, launch status, and upgrades are managed from the Agents page.
HHermes
Main recommendation for an always-on personal assistant.
- Chat, voice input, memory, modules, status checks, and one-click skill packs.
- Best for research, drafts, monitoring, reminders, and repeatable multi-step work.
OCOpenClaw
Optional second agent for tool-heavy workflows.
- Use it only after the capacity check confirms enough CPU, RAM, and disk for the selected setup.
- Modules can be installed, removed, and audited from the same marketplace flow.
One-click deploy route
- 1Select HermesOpenClaw can be added later if the selected setup has enough capacity.
- 2Pick placementPrice and required balance are shown before launch.
- 3Prepare assistantApiClaude prepares the assistant service, memory, status checks, and chat endpoint.
- 4Use and upgradeIf modules or load grow, the page suggests an upgrade before users feel it.
Models
Use the exact model id in Chat, IDEs, SDKs, and API calls. Paid routes and free promo aliases are listed separately.
Paid route IDs
Claude Fable 5
claude-fable-5
Claude Sonnet 5
claude-sonnet-5
Claude Opus 4.8
claude-opus-4-8
Claude Opus 4.7
claude-opus-4-7
Claude Sonnet 4.6
claude-sonnet-4-6
Claude Sonnet 4.5
claude-sonnet-4-5
Claude Opus 4.6
claude-opus-4-6
Claude Opus 4.5
claude-opus-4-5
Claude Haiku 4.5
claude-haiku-4-5
ChatGPT 5.5
gpt-5-5
GPT 5.4
gpt-5-4
Free promo route IDs
Free Opus 4.8PROMO
free-opus-4.8
Free Opus 4.7PROMO
free-opus-4.7
Free Opus 4.6PROMO
free-opus-4.6
Free Sonnet 5PROMO
free-sonnet-5
Free Sonnet 4.6PROMO
free-sonnet-4.6
Free Fable 5PROMO
free-fable-5
Free GPT 5.5PROMO
free-gpt-5.5
Free GPT 5.4PROMO
free-gpt-5.4
Billing price rows
Claude Opus 4.7
claude-opus-4-7
Claude Opus 4.8NEW
claude-opus-4-8
Claude Fable 5NEW
claude-fable-5
Claude Opus 4.6
claude-opus-4-6
ChatGPT 5.5NEW
gpt-5-5
GPT 5.4LOW COST
gpt-5-4
Claude Opus 4.5
claude-opus-4-5
Claude Sonnet 5NEW
claude-sonnet-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 million tokens. Checkout opens through the selected payment method and balance updates after confirmation.
Paid requests from Chat, API, Telegram, and AI assistants use the same model prices.
| Model | Input | Output | Approx. input | Approx. output |
|---|---|---|---|---|
Claude Opus 4.7claude-opus-4-7 |
40c | 40c | $0.40 | $0.40 |
Claude Opus 4.8 NEWclaude-opus-4-8 |
40c | 40c | $0.40 | $0.40 |
Claude Fable 5 NEWclaude-fable-5 |
80c | 80c | $0.80 | $0.80 |
Claude Opus 4.6claude-opus-4-6 |
40c | 40c | $0.40 | $0.40 |
ChatGPT 5.5 NEWgpt-5-5 |
40c | 40c | $0.40 | $0.40 |
GPT 5.4 LOW COSTgpt-5-4 |
19c | 19c | $0.19 | $0.19 |
Claude Opus 4.5claude-opus-4-5 |
40c | 40c | $0.40 | $0.40 |
Claude Sonnet 5 NEWclaude-sonnet-5 |
38c | 38c | $0.38 | $0.38 |
Claude Sonnet 4.6claude-sonnet-4-6 |
32c | 32c | $0.32 | $0.32 |
Claude Sonnet 4.5claude-sonnet-4-5 |
32c | 32c | $0.32 | $0.32 |
Claude Haiku 4.5claude-haiku-4-5 |
32c | 32c | $0.32 | $0.32 |
Included assistant models
These models are included for managed assistants while free capacity is available. Free Opus/Sonnet aliases may be limited by availability.
| Model | Scope | Input | Output |
|---|---|---|---|
Qwen3 235B AGENT FREEQwen/Qwen3-235B-A22B-Instruct-2507-FP8 |
Managed agents only | 0c | 0c |
Kimi K2.6 AGENT FREEmoonshotai/Kimi-K2.6 |
Managed agents only | 0c | 0c |
MiniMax M2.7 AGENT FREEMiniMaxAI/MiniMax-M2.7 |
Managed agents only | 0c | 0c |
Qwen3 235B Online AGENT FREEQwen/Qwen3-235B-A22B-Instruct-2507-FP8:online |
Managed agents only | 0c | 0c |
Kimi K2.6 Online AGENT FREEmoonshotai/Kimi-K2.6:online |
Managed agents only | 0c | 0c |
MiniMax M2.7 Online AGENT FREEMiniMaxAI/MiniMax-M2.7:online |
Managed agents only | 0c | 0c |
Free Opus 4.8 PROMO FREEfree-opus-4.8 |
Chat and agents when the promo pool is available | 0c | 0c |
Free Sonnet 5 PROMO FREEfree-sonnet-5 |
Chat and agents when the promo pool is available | 0c | 0c |
Free Fable 5 PROMO FREEfree-fable-5 |
Chat and agents when the promo pool is available | 0c | 0c |
Free Sonnet 4.6 PROMO FREEfree-sonnet-4.6 |
Chat and agents when the promo pool is available | 0c | 0c |
Free GPT 5.5 PROMO FREEfree-gpt-5.5 |
Chat and agents when the promo pool is available | 0c | 0c |
Free GPT 5.4 PROMO FREEfree-gpt-5.4 |
Chat and agents when the promo pool is available | 0c | 0c |
Free Opus 4.7 PROMO FREEfree-opus-4.7 |
Chat and agents when the promo pool is available | 0c | 0c |
Free Opus 4.6 PROMO FREEfree-opus-4.6 |
Chat and agents when the promo pool is available | 0c | 0c |
API
Use one ApiClaude key for OpenAI-compatible tools, Anthropic-style clients, web chat, and managed agents.
https://apiclaude.org/v1
POST
/chat/completions
OpenAI
curl https://apiclaude.org/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"claude-sonnet-4-6","messages":[{"role":"user","content":"Hi"}]}'POST
/messages
Anthropic
curl https://apiclaude.org/v1/messages \
-H "x-api-key: YOUR_API_KEY" \
-H "anthropic-version: 2023-06-01" \
-H "Content-Type: application/json" \
-d '{"model":"claude-sonnet-4-6","max_tokens":512,"messages":[{"role":"user","content":"Hi"}]}'RETRY
Transient retry policy
Use this in IDE agents and SDK clients
Retry short network failures and 408, 409, 425, 429, 500, 502, 503, 504. Do not retry 400, 401, 403, or validation errors without changing the request.
import time
from openai import OpenAI
client = OpenAI(
api_key="YOUR_API_KEY",
base_url="https://apiclaude.org/v1",
)
retry_status = {408, 409, 425, 429, 500, 502, 503, 504}
last_error = None
for attempt in range(3):
try:
response = client.chat.completions.create(
model="claude-sonnet-4-6",
messages=[{"role": "user", "content": "Hi"}],
max_tokens=512,
)
print(response.choices[0].message.content)
break
except Exception as exc:
last_error = exc
status = getattr(exc, "status_code", None)
if status not in retry_status and status is not None:
raise
if attempt == 2:
raise
time.sleep(0.5 * (2 ** attempt))
FAQ
Do I need a separate account for agents?
No. Agents use the same ApiClaude account, balance, API keys, and usage history.
Why can one IDE request cost more?
Coding agents often send chat history and project context. Use shorter sessions, summaries, and separate keys with limits.
Where is voice input available?
In the web chat and Hermes chat. Voice is transcribed first, then the text is sent through the selected model.
What should a new user open first?
Start with Chat if you want an answer now, IDE if you need coding tools, or Agents if you need long-running automation.