Have an LLM ask another LLM
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add ask-llm --env ASKLLM_CONFIG='<ASKLLM_CONFIG>' -- uvx --from git+https://github.com/joubertb/mcp-askllm mcp-askllm~/.codex/config.toml
[mcp_servers.ask-llm]
command = "uvx"
args = ["--from", "git+https://github.com/joubertb/mcp-askllm", "mcp-askllm"]
env = { ASKLLM_CONFIG = "<ASKLLM_CONFIG>" }
# codex mcp add ask-llm --env ASKLLM_CONFIG='<ASKLLM_CONFIG>' -- uvx --from git+https://github.com/joubertb/mcp-askllm mcp-askllm~/.cursor/mcp.json
{
"mcpServers": {
"ask-llm": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/joubertb/mcp-askllm",
"mcp-askllm"
],
"env": {
"ASKLLM_CONFIG": "<ASKLLM_CONFIG>"
}
}
}
}.vscode/mcp.json
{
"servers": {
"ask-llm": {
"type": "stdio",
"command": "uvx",
"args": [
"--from",
"git+https://github.com/joubertb/mcp-askllm",
"mcp-askllm"
],
"env": {
"ASKLLM_CONFIG": "<ASKLLM_CONFIG>"
}
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"ask-llm": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/joubertb/mcp-askllm",
"mcp-askllm"
],
"env": {
"ASKLLM_CONFIG": "<ASKLLM_CONFIG>"
}
}
}
}claude_desktop_config.json
{
"mcpServers": {
"ask-llm": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/joubertb/mcp-askllm",
"mcp-askllm"
],
"env": {
"ASKLLM_CONFIG": "<ASKLLM_CONFIG>"
}
}
}
}.trae/mcp.json
{
"mcpServers": {
"ask-llm": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/joubertb/mcp-askllm",
"mcp-askllm"
],
"env": {
"ASKLLM_CONFIG": "<ASKLLM_CONFIG>"
}
}
}
}Environment variables
ASKLLM_CONFIG
Replace the placeholders with your own credentials and keep them out of version control.