Claude API Base URL Configuration Guide for Cursor, Cline, Dify, Open WebUI, and More

Base URL mistakes are one of the most common reasons Claude API integrations fail.

The API key can be correct. The model name can be correct. But if the client expects an OpenAI-compatible base URL and you forget /v1, the request may fail with a 404. If the client expects an Anthropic-style base host and you add /v1 in the wrong field, the client may generate a duplicated path.

This guide explains what to enter for Cursor, Cline, Dify, ChatBox, Cherry Studio, Open WebUI, Claude Desktop-style clients, and command-line verification.

The short rule

Start with this table. In most cases, it is all you need.

The rule:

OpenAI-compatible tool      -> https://gw.claudeapi.com/v1
Anthropic-compatible client -> https://gw.claudeapi.com

OpenAI-compatible tool      -> https://gw.claudeapi.com/v1
Anthropic-compatible client -> https://gw.claudeapi.com

Do not add a trailing slash:

Correct: https://gw.claudeapi.com/v1
Avoid:   https://gw.claudeapi.com/v1/

Correct: https://gw.claudeapi.com/v1
Avoid:   https://gw.claudeapi.com/v1/

Some tools tolerate the trailing slash. Others concatenate paths poorly and produce double slashes or 404 errors.

Base URL is not the full endpoint

A Base URL field usually expects only the provider root plus API version:

https://gw.claudeapi.com/v1

https://gw.claudeapi.com/v1

The tool or SDK then appends the endpoint:

/chat/completions
/models

/chat/completions
/models

So the final OpenAI-compatible request becomes:

https://gw.claudeapi.com/v1/chat/completions

https://gw.claudeapi.com/v1/chat/completions

Do not paste this full endpoint into a field named Base URL, API Base, API Host, or API address:

Wrong for a Base URL field:
https://gw.claudeapi.com/v1/chat/completions

Wrong for a Base URL field:
https://gw.claudeapi.com/v1/chat/completions

That is the request endpoint, not the base.

Cursor

Use Cursor’s OpenAI-compatible or custom OpenAI base URL configuration if your Cursor version exposes it.

Typical setup:

1. Open Cursor settings.
2. Go to the Models or API provider section.
3. Select OpenAI Compatible or enable the custom OpenAI Base URL option.
4. Enter https://gw.claudeapi.com/v1.
5. Paste your ClaudeAPI key.
6. Enter the exact model ID.
7. Run Verify if Cursor provides a verification button.

Common failures:

Cline

Cline documents an OpenAI Compatible provider mode. Use that mode for ClaudeAPI’s OpenAI-compatible gateway.

Steps:

1. Open the Cline panel in VS Code.
2. Open Settings.
3. Select OpenAI Compatible as the provider.
4. Set Base URL to https://gw.claudeapi.com/v1.
5. Paste your ClaudeAPI key.
6. Enter the full model ID manually.
7. Save and send a short test message.

If the status bar still shows an old model, reload VS Code and check the selected provider again.

If you use the Claude Code VS Code plugin

Do not confuse Cline with the Claude Code VS Code plugin.

If your workflow uses Claude Code and CC Switch, the cleaner path is usually to let CC Switch manage the active provider configuration, then enable the option that applies the selected provider to the Claude Code plugin.

Recommended check:

After changing CC Switch settings, restart VS Code and send a test message from the Claude Code plugin.

Dify

In Dify, add ClaudeAPI as an OpenAI-compatible model provider.

Important details:

• Enter only https://gw.claudeapi.com/v1.
• Do not enter /chat/completions.
• Add model IDs manually if Dify does not auto-detect them.
• Enable the model before using it in a Dify app.

Successful setup means the model appears in Dify’s model list and can return a normal test response.

ChatBox

ChatBox can expose both native Claude/Anthropic and OpenAI-compatible modes. When using a ClaudeAPI key through the OpenAI-compatible gateway, choose the OpenAI mode.

If you select a native Claude API mode while using an OpenAI-compatible ClaudeAPI key, authentication or endpoint errors are likely.

Cherry Studio

Use the OpenAI Compatible provider type.

Typical setup:

1. Open Settings.
2. Go to model services or AI providers.
3. Add a provider.
4. Select OpenAI Compatible.
5. Enter https://gw.claudeapi.com/v1.
6. Paste your ClaudeAPI key.
7. Add the model ID in model management.
8. Return to chat and select the model.

Common issues:

Open WebUI

Open WebUI supports OpenAI-compatible connections through the admin settings.

Steps:

1. Log in as an administrator.
2. Open Admin Settings.
3. Go to Connections.
4. Add or edit the OpenAI-compatible connection.
5. Set the URL to https://gw.claudeapi.com/v1.
6. Paste the ClaudeAPI key.
7. Save and check whether models appear.

You can test model discovery with:

curl https://gw.claudeapi.com/v1/models \
  -H "Authorization: Bearer YOUR_CLAUDEAPI_KEY"

curl https://gw.claudeapi.com/v1/models \
  -H "Authorization: Bearer YOUR_CLAUDEAPI_KEY"

A successful response should contain an OpenAI-style model list:

{
  "object": "list",
  "data": [
    {
      "id": "claude-haiku-4-5-20251001",
      "object": "model"
    }
  ]
}

{
  "object": "list",
  "data": [
    {
      "id": "claude-haiku-4-5-20251001",
      "object": "model"
    }
  ]
}

If the model list is empty, check the Base URL, saved API key, model enablement, and admin permissions.

Claude Desktop and Anthropic-compatible clients

Claude Desktop requires extra caution because not every Claude Desktop version exposes a supported way to replace the inference endpoint. Some builds focus on Claude account sign-in and MCP server configuration rather than custom third-party model endpoints.

If your client or wrapper explicitly supports an Anthropic-compatible base URL, use the base host without /v1:

https://gw.claudeapi.com

https://gw.claudeapi.com

The client should append:

/v1/messages

/v1/messages

The final request becomes:

https://gw.claudeapi.com/v1/messages

https://gw.claudeapi.com/v1/messages

If you enter https://gw.claudeapi.com/v1 in a field where the client already appends /v1/messages, the final path may become malformed.

For Anthropic-compatible manual testing, use:

curl https://gw.claudeapi.com/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_CLAUDEAPI_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-haiku-4-5-20251001",
    "max_tokens": 100,
    "messages": [{"role": "user", "content": "Reply with OK only."}]
  }'

curl https://gw.claudeapi.com/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_CLAUDEAPI_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-haiku-4-5-20251001",
    "max_tokens": 100,
    "messages": [{"role": "user", "content": "Reply with OK only."}]
  }'

Expected result:

• The response contains a content array.
• content[0].text contains the model reply.

If your Claude Desktop build does not expose a custom API base field, do not force this configuration into unrelated MCP settings. Use ClaudeAPI with tools that explicitly support custom OpenAI-compatible or Anthropic-compatible endpoints.

Troubleshooting flow

When a tool fails, test the API outside the tool first. This separates provider, network, and model issues from UI configuration issues.

Step 1: Test OpenAI-compatible chat completions

Use this for Cursor, Cline, Dify, ChatBox, Cherry Studio, and Open WebUI:

curl https://gw.claudeapi.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_CLAUDEAPI_KEY" \
  -d '{
    "model": "claude-haiku-4-5-20251001",
    "messages": [{"role": "user", "content": "Reply with OK only."}],
    "max_tokens": 16
  }'

curl https://gw.claudeapi.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_CLAUDEAPI_KEY" \
  -d '{
    "model": "claude-haiku-4-5-20251001",
    "messages": [{"role": "user", "content": "Reply with OK only."}],
    "max_tokens": 16
  }'

Expected result:

• The response contains a choices array.
• choices[0].message.content contains text.
• The usage field may show token counts.

Step 2: Confirm the API key source

Use a key generated in the ClaudeAPI console. Keys from Anthropic or another provider are not interchangeable with ClaudeAPI keys.

Also check:

• No leading or trailing spaces
• No missing characters after copy/paste
• The key has available balance or quota
• The target model is enabled for the account

Step 3: Check the URL pattern

Error reference

FAQ

Should the Base URL end with a slash?

No. Use:

https://gw.claudeapi.com/v1

https://gw.claudeapi.com/v1

not:

https://gw.claudeapi.com/v1/

https://gw.claudeapi.com/v1/

Why do OpenAI-compatible tools need /v1?

OpenAI-compatible tools usually expect the base URL to include the API version. They then append paths such as /chat/completions or /models.

Why do Anthropic-compatible clients use the host without /v1?

Anthropic-style clients often append /v1/messages themselves. If you include /v1 in the base field, the final path may be duplicated or invalid.

Is a ClaudeAPI key the same as an Anthropic key?

No. Use the key generated in the ClaudeAPI console. Provider keys are not interchangeable.

Can I use the same ClaudeAPI key in multiple tools?

Yes. One key can be used in multiple clients, and usage is billed according to requests. For team environments, separate keys per tool, user, or environment make auditing easier.

Where do I find model names?

Check the ClaudeAPI model page or console and copy the exact model ID. Common examples include:

claude-haiku-4-5-20251001
claude-sonnet-4-6
claude-opus-4-8

claude-haiku-4-5-20251001
claude-sonnet-4-6
claude-opus-4-8

A tool says the connection succeeded, but chat still fails. Why?

Some connection checks only validate authentication or model discovery. A real chat request also depends on the selected model, enabled capabilities, quota, and request format. Run the curl chat completion test to isolate the issue.

Final checklist

Before debugging anything else, verify these items:

• [ ] You selected OpenAI Compatible mode for OpenAI-style tools.
• [ ] The Base URL is https://gw.claudeapi.com/v1 for OpenAI-compatible tools.
• [ ] The Base URL is https://gw.claudeapi.com only for Anthropic-compatible base-host fields.
• [ ] The field contains a base path, not /chat/completions or /messages.
• [ ] The URL has no trailing slash.
• [ ] The key comes from ClaudeAPI.
• [ ] The model ID is copied exactly from the console.
• [ ] The model is enabled and the account has available quota.

Once you remember the rule, new tools are easier to configure: first identify whether the tool speaks OpenAI-compatible or Anthropic-compatible API, then choose the matching Base URL.

Related guides

• Migrate from OpenAI API to Claude API
• Continue VS Code Claude API setup guide
• Claude API pricing and model selection guide
• Claude API cost estimation and budget guide

Sources

• Cline Docs: OpenAI Compatible provider
• Dify Marketplace: OpenAI-API-compatible provider
• Open WebUI Docs: OpenAI-compatible provider connection
• Anthropic Docs: Messages API
• Vercel AI SDK: OpenAI Compatible Providers

Disclosure

This article was translated and adapted from a Chinese ClaudeAPI draft with AI assistance, then reviewed against public documentation available on July 14, 2026. ClaudeAPI is an independent third-party technical service provider. It is not affiliated with Anthropic and does not guarantee the availability of Anthropic services.