Supported LLM Providers in OpenClaude: Complete Provider List and Configuration Guide

OpenClaude supports 12 LLM providers including Anthropic's official API, Amazon Bedrock, Google Vertex AI, OpenAI, GitHub Copilot, and NVIDIA NIM.

The OpenClaude CLI abstracts multiple language model services behind a unified API provider architecture. This guide covers every supported provider, how to configure each one, and the source code implementation that powers provider selection.

Complete List of Supported LLM Providers

OpenClaude recognizes the following provider identifiers, defined in the LegacyAPIProvider type at src/utils/model/providers.ts (lines 10-27):

Provider ID	Service
firstParty	Anthropic's official API (default)
bedrock	Amazon Bedrock
vertex	Google Vertex AI
foundry	Foundry (internal/experimental)
openai	OpenAI and OpenAI-compatible endpoints
gemini	Google Gemini
github	GitHub Copilot API
codex	OpenAI Codex (legacy)
nvidia-nim	NVIDIA NIM
minimax	Minimax AI
mistral	Mistral AI
xai	XAI
xiaomi-mimo	Xiaomi MIMO

The LegacyAPIProvider type serves as the canonical registry of all supported LLM providers in the OpenClaude codebase.

How OpenClaude Selects the Active Provider

Provider selection happens at runtime through the getAPIProvider() function (lines 32-81 in src/utils/model/providers.ts). This function evaluates environment variables and route metadata to determine which backend to use.

Key environment variables that control provider selection:

• CLAUDE_CODE_USE_FOUNDRY — enables the Foundry provider
• CLAUDE_CODE_USE_GITHUB — switches to GitHub Copilot
• NVIDIA_NIM — activates NVIDIA NIM inference
• AZURE_OPENAI_API_KEY or OPENAI_API_KEY — triggers OpenAI provider detection

Code Examples for Provider Management

Resolve the Active Provider

import { getAPIProvider } from './src/utils/model/providers.js';

const provider = getAPIProvider();
console.log(`OpenClaude is using the "${provider}" provider`);
// Output: "OpenClaude is using the "openai" provider"

Check for Anthropic First-Party Mode

import { isFirstPartyAnthropicProvider } from './src/utils/model/providers.js';

if (isFirstPartyAnthropicProvider()) {
  console.log('Running against Anthropic's official API');
}

Switch to GitHub Copilot Provider

process.env.CLAUDE_CODE_USE_GITHUB = '1';
process.env.OPENAI_MODEL = 'claude-sonnet-3.5';

import { getAPIProvider } from './src/utils/model/providers.js';

console.log(getAPIProvider()); // → "github"

Key Source Files for Provider Architecture

File	Purpose
src/utils/model/providers.ts	Central definition of all supported LLM providers and selection logic
src/integrations/routeMetadata.ts	Route-to-provider mapping for getAPIProvider()
src/utils/providerProfiles.ts	Saved provider profiles and custom endpoints
src/services/api/providerConfig.ts	Provider-specific configuration (tokens, base URLs, timeouts)

These four files implement OpenClaude's multi-provider abstraction layer.

Provider-Specific Configuration Notes

First-party Anthropic requires ANTHROPIC_API_KEY and uses https://api.anthropic.com by default.

Amazon Bedrock needs AWS credentials (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY) plus AWS_REGION set to a Bedrock-supported region.

GitHub Copilot supports an optional Anthropic native mode when OPENAI_MODEL is set to a Claude model identifier.

NVIDIA NIM expects NVIDIA_NIM environment variable pointing to your NIM inference endpoint.

Summary

• OpenClaude supports 12 LLM providers ranging from commercial APIs to self-hosted inference.
• Provider IDs are defined in src/utils/model/providers.ts and selected via getAPIProvider().
• Runtime configuration uses environment variables rather than static config files.
• The architecture allows seamless switching between providers without code changes.

Frequently Asked Questions

How do I switch from Anthropic to OpenAI in OpenClaude?

Set OPENAI_API_KEY in your environment and ensure ANTHROPIC_API_KEY is unset. The getAPIProvider() function automatically detects OpenAI credentials and returns "openai". For Azure OpenAI, use AZURE_OPENAI_API_KEY instead.

What is the default LLM provider in OpenClaude?

The first-party Anthropic provider (firstParty) is the default when ANTHROPIC_API_KEY is present and no override environment variables are set. This targets api.anthropic.com directly without intermediary services.

Can I use multiple providers simultaneously in one OpenClaude session?

No — OpenClaude uses a single active provider per process determined at startup by getAPIProvider(). To use different providers, launch separate shell sessions with distinct environment configurations or use provider profiles defined in src/utils/providerProfiles.ts.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}