Skip to main content
Glama

Gemini Context MCP Server

by ogoldberg
mcp-manifest.json8.38 kB
{ "name": "gemini-context", "version": "1.0.0", "description": "MCP server for adding Gemini context management with both session-based context and API caching", "capabilities": { "tools": true }, "tools": [ { "name": "generate_text", "description": "Generate text using Gemini with session-based context management", "parameters": { "sessionId": { "type": "string", "description": "Unique identifier for the conversation session" }, "message": { "type": "string", "description": "The user's message to process" } }, "examples": [ { "parameters": { "sessionId": "user-123", "message": "What is machine learning?" }, "description": "Basic question in a new session" }, { "parameters": { "sessionId": "user-123", "message": "What are some common algorithms used for it?" }, "description": "Follow-up question in the same session" } ] }, { "name": "get_context", "description": "Retrieve the current context for a session including all messages", "parameters": { "sessionId": { "type": "string", "description": "Unique identifier for the conversation session" } }, "examples": [ { "parameters": { "sessionId": "user-123" }, "description": "Get all context in a session" } ] }, { "name": "clear_context", "description": "Clear the context for a session, removing all stored messages", "parameters": { "sessionId": { "type": "string", "description": "Unique identifier for the conversation session" } }, "examples": [ { "parameters": { "sessionId": "user-123" }, "description": "Clear all conversation history in a session" } ] }, { "name": "add_context", "description": "Add a new entry to the context without generating a response", "parameters": { "content": { "type": "string", "description": "The content to add to context" }, "role": { "type": "string", "enum": ["user", "assistant", "system"], "description": "Role of the context entry" }, "metadata": { "type": "object", "properties": { "topic": { "type": "string", "description": "Topic for context organization" }, "tags": { "type": "array", "items": { "type": "string" }, "description": "Tags for context categorization" } }, "description": "Metadata for context tracking" } }, "examples": [ { "parameters": { "content": "I have a cat named Whiskers.", "role": "user", "metadata": { "topic": "pets", "tags": ["cat", "personal"] } }, "description": "Add user information with metadata" }, { "parameters": { "content": "The user should be given concise responses.", "role": "system", "metadata": { "topic": "preferences" } }, "description": "Add system instruction" } ] }, { "name": "search_context", "description": "Search for relevant context using semantic similarity", "parameters": { "query": { "type": "string", "description": "The search query to find relevant context" }, "limit": { "type": "number", "description": "Maximum number of context entries to return" } }, "examples": [ { "parameters": { "query": "pets" }, "description": "Find all context entries related to pets" }, { "parameters": { "query": "financial data", "limit": 5 }, "description": "Find up to 5 entries related to financial data" } ] }, { "name": "mcp_gemini_context_create_cache", "description": "Create a cache for frequently used large contexts for API-level caching (min 32K tokens recommended)", "parameters": { "displayName": { "type": "string", "description": "Friendly name for the cache" }, "content": { "type": "string", "description": "Large context to cache (system instructions, documents, etc)" }, "ttlSeconds": { "type": "number", "description": "Time to live in seconds (default: 3600)" } }, "examples": [ { "parameters": { "displayName": "Financial Analysis System", "content": "You are a specialized AI assistant for analyzing financial data...", "ttlSeconds": 3600 }, "description": "Create a cache for financial analysis prompts" } ] }, { "name": "mcp_gemini_context_generate_with_cache", "description": "Generate content using a cached context for cost optimization", "parameters": { "cacheName": { "type": "string", "description": "The cache name/ID from createCache" }, "userPrompt": { "type": "string", "description": "The user prompt to append to the cached context" } }, "examples": [ { "parameters": { "cacheName": "abc123", "userPrompt": "Explain what a P/E ratio is in simple terms." }, "description": "Generate a response using a cached financial context" } ] }, { "name": "mcp_gemini_context_list_caches", "description": "List all available caches", "parameters": {}, "examples": [ { "parameters": {}, "description": "List all available caches" } ] }, { "name": "mcp_gemini_context_update_cache_ttl", "description": "Updates a cache's TTL (time to live)", "parameters": { "cacheName": { "type": "string", "description": "Cache name/ID" }, "ttlSeconds": { "type": "number", "description": "New TTL in seconds" } }, "examples": [ { "parameters": { "cacheName": "abc123", "ttlSeconds": 7200 }, "description": "Extend cache TTL to 2 hours" } ] }, { "name": "mcp_gemini_context_delete_cache", "description": "Deletes a cache", "parameters": { "cacheName": { "type": "string", "description": "Cache name/ID" } }, "examples": [ { "parameters": { "cacheName": "abc123" }, "description": "Delete a cache that is no longer needed" } ] } ], "usage": { "context_management": { "description": "Session-based context management for conversations", "workflow": [ "1. Start with generate_text providing a sessionId", "2. Continue the conversation using the same sessionId", "3. Use get_context to retrieve conversation history", "4. Use clear_context when starting a new topic" ] }, "api_caching": { "description": "API-level caching for cost optimization with large contexts", "workflow": [ "1. Create a cache with mcp_gemini_context_create_cache", "2. Generate responses using mcp_gemini_context_generate_with_cache", "3. Update TTL with mcp_gemini_context_update_cache_ttl if needed", "4. Delete cache with mcp_gemini_context_delete_cache when done" ], "requirements": [ "Minimum context size of 32K tokens recommended for cost benefits", "Must use a stable model version (e.g., gemini-1.5-pro-001)" ] } } }

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/ogoldberg/gemini-context-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server