
How to add Outsource MCP to Claude Desktop
Give your AI assistant its own AI assistants — outsource text and image generation to 20+ model providers. Paste the config into ~/Library/Application Support/Claude/claude_desktop_config.json and restart Claude Desktop.
Last updated June 14, 2026 · 30★ · stdio · apikey
Claude Desktop config for Outsource MCP
uvx --from git+https://github.com/gwbischof/outsource-mcp.git outsource-mcp{
"mcpServers": {
"outsource-mcp": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/gwbischof/outsource-mcp.git",
"outsource-mcp"
],
"env": {
"OPENAI_API_KEY": "your-openai-key",
"ANTHROPIC_API_KEY": "your-anthropic-key",
"GOOGLE_API_KEY": "your-google-key",
"GROQ_API_KEY": "your-groq-key",
"DEEPSEEK_API_KEY": "your-deepseek-key",
"XAI_API_KEY": "your-xai-key",
"PERPLEXITY_API_KEY": "your-perplexity-key",
"COHERE_API_KEY": "your-cohere-key",
"FIREWORKS_API_KEY": "your-fireworks-key",
"HUGGINGFACE_API_KEY": "your-huggingface-key",
"MISTRAL_API_KEY": "your-mistral-key",
"NVIDIA_API_KEY": "your-nvidia-key",
"OLLAMA_HOST": "http://localhost:11434",
"OPENROUTER_API_KEY": "your-openrouter-key",
"TOGETHER_API_KEY": "your-together-key",
"CEREBRAS_API_KEY": "your-cerebras-key",
"DEEPINFRA_API_KEY": "your-deepinfra-key",
"SAMBANOVA_API_KEY": "your-sambanova-key"
}
}
}
}Requires `uv` (the Python package runner). Install it from https://docs.astral.sh/uv/ if `uvx` is not found.
Setup steps
- 1Open Claude Desktop → Settings → Developer → Edit Config (this opens ~/Library/Application Support/Claude/claude_desktop_config.json).
- 2Paste the Outsource MCP config below under the top-level "mcpServers" key.
- 3Fill in any placeholder secrets (API keys, paths) in the snippet.
- 4Save the file, then fully quit and reopen Claude Desktop.
- 5Open a chat and confirm Outsource MCP's tools appear under the 🔌 tools menu.
Before you start
- Python 3.11 or higher
- uv package manager (provides uvx)
- At least one model provider API key (e.g. OPENAI_API_KEY) for the providers you want to use
What Outsource MCP can do in Claude Desktop
outsource_textCreates an Agno agent with the specified provider and model to generate a text response. Arguments: provider (e.g. "openai", "anthropic", "google", "groq"), model (e.g. "gpt-4o", "claude-3-5-sonnet-20241022", "gemini-2.0-flash-exp"), and prompt (the text to send).
outsource_imageGenerates an image using an AI model and returns the URL of the generated image. Arguments: provider (currently only "openai" is supported), model ("dall-e-3" or "dall-e-2"), and prompt (the image generation prompt).
Security
Requires API keys for whichever model providers you enable; each key is passed via environment variables in the MCP client config. Only configure keys for providers you intend to use. Outbound calls go to third-party model provider APIs (and incur their usage costs).
Outsource MCP + Claude Desktop FAQ
Where is the Claude Desktop config file?
Claude Desktop reads MCP servers from ~/Library/Application Support/Claude/claude_desktop_config.json. Paste the Outsource MCP config there under the "mcpServers" key and restart the client.
Is Outsource MCP safe to use with Claude Desktop?
Requires API keys for whichever model providers you enable; each key is passed via environment variables in the MCP client config. Only configure keys for providers you intend to use. Outbound calls go to third-party model provider APIs (and incur their usage costs).
Which providers are supported?
20+ providers including OpenAI, Anthropic, Google, Groq, DeepSeek, xAI, Perplexity, Cohere, Mistral, NVIDIA, HuggingFace, Ollama, Fireworks, OpenRouter, Together, Cerebras, DeepInfra, SambaNova, plus enterprise providers (AWS Bedrock, Azure AI, IBM WatsonX, LiteLLM, Vercel v0, Meta Llama). Provider names are case-insensitive.
Can it generate images with providers other than OpenAI?
No. Image generation is currently only supported by OpenAI's DALL-E 2 and DALL-E 3 models. All other providers support text generation only.
Do I need to configure every API key in the config?
No. The environment variables are optional — only include API keys for the providers you actually plan to use.