Context

Provides persistent vector-based context management for AI agents with semantic search, batch operations, and metadata filtering using Upstash Vector DB and Google AI.

typescript

GitHub Stars

typescript

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "raunak-dev-18-context-mcp": {
      "command": "node",
      "args": [
        "path/to/context-mcp/dist/index.js"
      ],
      "env": {
        "GOOGLE_AI_API_KEY": "your_key",
        "UPSTASH_VECTOR_REST_URL": "your_url",
        "UPSTASH_VECTOR_REST_TOKEN": "your_token"
      }
    }
  }
}

You can run a dedicated MCP server that manages persistent context for AI agents, enabling semantic searches, batch context operations, and metadata filtering. This server stores vectors in Upstash Vector DB and uses Google AI for embeddings, making it easy to add, query, and manage contextual information for your AI workflows.

How to use

You will connect your MCP client to a local MCP server that runs as a standard Node process. The server exposes functions to add context, batch add contexts, query for relevant context, delete single or multiple contexts, and fetch statistics. Start the server in the background of your AI agent workflow, then call the available endpoints from your agent configuration to enrich conversations, searches, and project notes with semantic awareness.

How to install

# Clone the repository
git clone <your-repo-url>
cd context-mcp

# Install dependencies
npm install

# Build the project
npm run build

Configuration and prerequisites

Before running the MCP server, supply credentials for storage and embeddings. You will configure environment variables in a .env file that the server reads on startup.

UPSTASH_VECTOR_REST_URL=your_upstash_vector_url
UPSTASH_VECTOR_REST_TOKEN=your_upstash_vector_token
GOOGLE_AI_API_KEY=your_google_ai_api_key

Available tools

add_context

Store a single piece of context with a unique id, content, and optional metadata for filtering.

add_contexts_batch

Store multiple contexts efficiently in a single operation, reducing round-trips.

query_context

Perform a semantic search to retrieve contexts related to a natural language query, with optional topK and filter parameters.

delete_context

Remove a single context by its id.

delete_contexts_batch

Delete multiple contexts by their ids in a single operation.

get_stats

Retrieve database statistics such as vector count and dimensions.

Built by

VeilStrat

AI signals for GTM teams