gemini-researcher MCP
A stateless MCP server that proxies research queries to Gemini CLI, allowing agents like Claude Code and Cursor to perform deep codebase analysis locally. It reduces the calling agent's context window and model usage by offloading heavy analysis to Gemini's free tier, returning structured JSON results while maintaining read-only safety.
Package and access
Local package with source-attributed installation metadata. Koskamo does not execute packages during review.
What you can do with it
- Offload deep repository analysis to Gemini CLI
- Reduce agent context usage by avoiding file copying
- Perform cost-optimized codebase research
- Analyze large codebases without burning main model quota
Source and maintenance
- Maintenance checked
Packages and installation
Source-provided package data. Koskamo has not executed or live-probed this package.
gemini-researcher 1.0.2
{
"version": "1.0.2",
"transport": "stdio",
"identifier": "gemini-researcher",
"registryType": "npm",
"environmentVariables": [
{
"name": "GEMINI_API_KEY",
"format": "string",
"description": "Gemini API key (optional if you already authenticated Gemini CLI via \"gemini\" login)"
},
{
"name": "PROJECT_ROOT",
"format": "string",
"description": "Override the project root directory used for path validation (defaults to current working directory)"
},
{
"name": "RESPONSE_CHUNK_SIZE_KB",
"format": "string",
"description": "Chunk size threshold (KB) for large responses (default: 10)"
},
{
"name": "CACHE_TTL_MS",
"format": "string",
"description": "Chunk cache TTL in milliseconds (default: 3600000 / 1 hour)"
},
{
"name": "DEBUG",
"format": "string",
"description": "Enable debug logging (set to \"true\" or \"1\")"
},
{
"name": "GOOGLE_APPLICATION_CREDENTIALS",
"format": "string",
"description": "Vertex AI / Google auth: path to a service account JSON credentials file"
},
{
"name": "GOOGLE_CLOUD_PROJECT",
"format": "string",
"description": "Vertex AI / Google auth: GCP project ID (used by some auth configurations)"
},
{
"name": "VERTEX_AI_PROJECT",
"format": "string",
"description": "Vertex AI / Google auth: Vertex AI project identifier (used by some auth configurations)"
}
]
}Tools
Reported by the attributed source; not independently observed by Koskamo.
Connect from your client
Each snippet configures the same source-provided package; replace placeholders before use.
Cursor
{
"mcpServers": {
"gemini-researcher-mcp": {
"command": "npx",
"args": [
"-y",
"gemini-researcher@1.0.2"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"PROJECT_ROOT": "<YOUR_PROJECT_ROOT>",
"RESPONSE_CHUNK_SIZE_KB": "<YOUR_RESPONSE_CHUNK_SIZE_KB>",
"CACHE_TTL_MS": "<YOUR_CACHE_TTL_MS>",
"DEBUG": "<YOUR_DEBUG>",
"GOOGLE_APPLICATION_CREDENTIALS": "<YOUR_GOOGLE_APPLICATION_CREDENTIALS>",
"GOOGLE_CLOUD_PROJECT": "<YOUR_GOOGLE_CLOUD_PROJECT>",
"VERTEX_AI_PROJECT": "<YOUR_VERTEX_AI_PROJECT>"
}
}
}
}Cline
{
"mcpServers": {
"gemini-researcher-mcp": {
"command": "npx",
"args": [
"-y",
"gemini-researcher@1.0.2"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"PROJECT_ROOT": "<YOUR_PROJECT_ROOT>",
"RESPONSE_CHUNK_SIZE_KB": "<YOUR_RESPONSE_CHUNK_SIZE_KB>",
"CACHE_TTL_MS": "<YOUR_CACHE_TTL_MS>",
"DEBUG": "<YOUR_DEBUG>",
"GOOGLE_APPLICATION_CREDENTIALS": "<YOUR_GOOGLE_APPLICATION_CREDENTIALS>",
"GOOGLE_CLOUD_PROJECT": "<YOUR_GOOGLE_CLOUD_PROJECT>",
"VERTEX_AI_PROJECT": "<YOUR_VERTEX_AI_PROJECT>"
}
}
}
}VS Code
{
"servers": {
"gemini-researcher-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"gemini-researcher@1.0.2"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"PROJECT_ROOT": "<YOUR_PROJECT_ROOT>",
"RESPONSE_CHUNK_SIZE_KB": "<YOUR_RESPONSE_CHUNK_SIZE_KB>",
"CACHE_TTL_MS": "<YOUR_CACHE_TTL_MS>",
"DEBUG": "<YOUR_DEBUG>",
"GOOGLE_APPLICATION_CREDENTIALS": "<YOUR_GOOGLE_APPLICATION_CREDENTIALS>",
"GOOGLE_CLOUD_PROJECT": "<YOUR_GOOGLE_CLOUD_PROJECT>",
"VERTEX_AI_PROJECT": "<YOUR_VERTEX_AI_PROJECT>"
}
}
}
}Claude Desktop
{
"mcpServers": {
"gemini-researcher-mcp": {
"command": "npx",
"args": [
"-y",
"gemini-researcher@1.0.2"
],
"env": {
"GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
"PROJECT_ROOT": "<YOUR_PROJECT_ROOT>",
"RESPONSE_CHUNK_SIZE_KB": "<YOUR_RESPONSE_CHUNK_SIZE_KB>",
"CACHE_TTL_MS": "<YOUR_CACHE_TTL_MS>",
"DEBUG": "<YOUR_DEBUG>",
"GOOGLE_APPLICATION_CREDENTIALS": "<YOUR_GOOGLE_APPLICATION_CREDENTIALS>",
"GOOGLE_CLOUD_PROJECT": "<YOUR_GOOGLE_CLOUD_PROJECT>",
"VERTEX_AI_PROJECT": "<YOUR_VERTEX_AI_PROJECT>"
}
}
}
}Let your agent find its tools
Find servers like gemini-researcher MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.