gemini researcher for VS Code
io.github.capybearista/gemini-researcher
Stateless MCP server that proxies research queries to Gemini CLI, reducing agent context/model usage
client:VS Code
transport:stdio
runtime:npm
Install gemini researcher in VS Code
.vscode/mcp.json
{
"inputs": [
{
"type": "promptString",
"id": "gemini-api-key",
"description": "GEMINI_API_KEY",
"password": true
},
{
"type": "promptString",
"id": "project-root",
"description": "PROJECT_ROOT",
"password": true
},
{
"type": "promptString",
"id": "response-chunk-size-kb",
"description": "RESPONSE_CHUNK_SIZE_KB",
"password": true
},
{
"type": "promptString",
"id": "cache-ttl-ms",
"description": "CACHE_TTL_MS",
"password": true
},
{
"type": "promptString",
"id": "debug",
"description": "DEBUG",
"password": true
},
{
"type": "promptString",
"id": "google-application-credentials",
"description": "GOOGLE_APPLICATION_CREDENTIALS",
"password": true
},
{
"type": "promptString",
"id": "google-cloud-project",
"description": "GOOGLE_CLOUD_PROJECT",
"password": true
},
{
"type": "promptString",
"id": "vertex-ai-project",
"description": "VERTEX_AI_PROJECT",
"password": true
}
],
"servers": {
"gemini-researcher": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"gemini-researcher"
],
"env": {
"GEMINI_API_KEY": "${input:gemini-api-key}",
"PROJECT_ROOT": "${input:project-root}",
"RESPONSE_CHUNK_SIZE_KB": "${input:response-chunk-size-kb}",
"CACHE_TTL_MS": "${input:cache-ttl-ms}",
"DEBUG": "${input:debug}",
"GOOGLE_APPLICATION_CREDENTIALS": "${input:google-application-credentials}",
"GOOGLE_CLOUD_PROJECT": "${input:google-cloud-project}",
"VERTEX_AI_PROJECT": "${input:vertex-ai-project}"
}
}
}
}
VS Code prompts for each input the first time the server starts, so no secret is written to the file. VS Code docs