UseCortex MCP Server

Give your AI coding agent a persistent memory — read and write knowledge from any tool.

Website · Quick Start · Tools · Pricing

UseCortex MCP is a Model Context Protocol server that gives AI coding agents persistent memory. Your agent can read knowledge you've stored and write new discoveries back — all through a single encrypted endpoint.

No context window limits. No copy-pasting. No outdated docs. Your AI remembers everything.

How It Works

1. You chat — describe your coding standards, architecture decisions, client requirements, anything
2. UseCortex extracts — AI automatically organizes facts into structured topics
3. Your agent recalls — any MCP-compatible tool can query your knowledge instantly
4. Knowledge grows — your agent can push new discoveries back, so your knowledge base stays current

Features

• Two-way knowledge flow — read context into projects, write discoveries back
• Works with any MCP client — compatible with any AI coding agent that supports the Model Context Protocol
• Encrypted at rest — AES-256 encryption with per-user keys
• No vendor lock-in — your knowledge works across every AI tool
• Structured by topic — knowledge auto-organized, not dumped in one pile

Quick Start

Prerequisites

• A free UseCortex account
• An MCP-compatible AI coding agent

1. Generate an API key

Sign up at usecortex.net, then navigate to Settings → API Keys → Generate.

Note: Copy the key immediately after generation. It will not be shown again.

2. Install (one command)

Run this in your terminal — replace YOUR_API_KEY with the key from step 1:

claude mcp add usecortex --transport url https://api.usecortex.net/mcp --header "Authorization: Bearer YOUR_API_KEY"

That's it. Restart your AI coding agent and UseCortex tools are ready.

{
  "mcpServers": {
    "usecortex": {
      "type": "url",
      "url": "https://api.usecortex.net/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Available Tools

query_knowledge

Ask a natural language question and get an AI-powered answer based on your stored knowledge.

list_topics

List all knowledge topics. Takes no parameters.

add_knowledge

Write new knowledge to your base. Your agent can store discoveries, decisions, or patterns it finds.

get_topic

Retrieve all knowledge entries for a specific topic.

search_knowledge

Search knowledge entries by keyword (text match).

capture_session (Memory plan)

Capture an AI session summary into persistent memory. Store what you learned, decided, or built.

recall_memory (Memory plan)

Search across all captured session memories using AI. Ask what you worked on or what decisions were made.

list_sessions (Memory plan)

List recent captured session memories.

Use Cases

Pricing

Manage your plan at usecortex.net.

Security

• All communication encrypted over HTTPS
• Knowledge encrypted at rest with AES-256
• Per-user encryption keys — even we can't read your data
• API keys scoped per user with instant revocation

Contributing

Found a bug or have a feature request? Open an issue.

Links

• UseCortex — Create your account and manage knowledge
• cortex-cli — CLI for markdown sync & terminal usage
• Model Context Protocol — Learn about MCP

License

MIT