{
  "markdown": "# Cortex — Semantic Memory for AI Agents\n\n> One external brain, model-agnostic, in production. Connect it to Claude in two minutes — free trial, no credit card.\n\n**Cortex** is a patent-pending semantic memory infrastructure developed by [SKYNETLAB](https://skynetlab.net/) (Bergamo, Italy). Connected to your AI assistant as a remote MCP connector, it lets your conversations **save decisions, context and sources — and find them again in every later session**. Your AI remembers, and can show *why* it remembers.\n\n🌐 Website: [skynetlab-cortex.com](https://skynetlab-cortex.com/) · 📄 [Patent](https://skynetlab-cortex.com/brevetto) · 📊 [Benchmark](https://skynetlab-cortex.com/benchmark) · 📚 [Paper](https://skynetlab-cortex.com/paper) · 💶 [Pricing](https://skynetlab-cortex.com/abbonamenti)\n\n[![MCP Registry](https://img.shields.io/badge/MCP%20Registry-io.github.FilippoPilo%2Fcortex-b8523d)](https://registry.modelcontextprotocol.io) [![Listed on mcpservers.org](https://mcpservers.org/badge.svg)](https://mcpservers.org/servers/filippopilo/cortex-connector)\n\n> 🇮🇹 Versione italiana: [README.it.md](README.it.md)\n\n![Cortex — one external brain, shared by all your AIs](assets/cortex-site-hero.png)\n\n---\n\n## Why Cortex is not \"just another memory plugin\"\n\nMost memory tools for AI agents record what the agent did and retrieve similar text later. Cortex models **what is true and how reliable it is**:\n\n| Capability | What it means |\n|---|---|\n| **Quality Gate on write** | Every memory passes a novelty/redundancy check before being stored. No junk accumulation. |\n| **Typed claims & conflict tracking** | Facts are extracted as claims; contradictions between memories are detected and tracked, not silently overwritten. |\n| **Knowledge graph** | Memories are linked by entities and typed relations, not just vector similarity. |\n| **Coherence metric (Ψ_C)** | Each memory carries a native coherence score — the system knows how well a memory fits what it already knows. |\n| **Memory consolidation (\"REM\")** | A continuous background cycle consolidates episodic memories into long-term knowledge. |\n| **Transparent evidence** | Answers can cite the memories and sources they come from. |\n\nCortex runs on globally distributed edge infrastructure operated in the **European Union**. The memory engine is covered by Italian patent application **UIBM 102026000014026** (filed 15 May 2026).\n\n## Quick start — Claude (web / desktop / mobile)\n\nRequirements: a Claude plan that supports custom connectors (the \"Connectors\" entry appears in Settings). No Cortex account needed beforehand — it is created on first sign-in.\n\n1. In Claude open **Settings → Connectors → Add custom connector**\n2. Paste this URL and confirm:\n\n   ```\n   https://skynetlab-cortex-saas-mcp.cortex-320.workers.dev/mcp\n   ```\n\n3. Authorize access on the Cortex consent page (Google or email sign-in). Your personal memory space is created automatically, isolated from every other user.\n\nThat's it. Talk to Claude as usual: *\"Remember that we chose supplier X\"*, *\"what did we decide about the budget?\"*, *\"show me the sources for this claim\"*. Claude picks the right tool on its own.\n\n## Quick start — Claude Code\n\n```bash\nclaude mcp add --transport http cortex https://skynetlab-cortex-saas-mcp.cortex-320.workers.dev/mcp\n```\n\nThen authenticate when prompted (`/mcp` shows connection status).\n\n## Quick start — other MCP clients (stdio bridge)\n\nFor MCP clients that don't support remote connectors natively, use the [`mcp-remote`](https://www.npmjs.com/package/mcp-remote) bridge. Example configuration (see [`examples/`](examples/)):\n\n```json\n{\n  \"mcpServers\": {\n    \"cortex\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"mcp-remote\", \"https://skynetlab-cortex-saas-mcp.cortex-320.workers.dev/mcp\"]\n    }\n  }\n}\n```\n\n## Quick start — this repository as a local stdio server\n\nThis repository also ships a small **stdio MCP server** (`server/`) that exposes the Cortex tools locally and forwards every call to the hosted service. It is meant for MCP clients, registries and evaluators that need a local process: it starts, lists its seven tools and answers introspection with no configuration at all; tool calls are forwarded when an access token is set.\n\n```bash\ngit clone https://github.com/FilippoPilo/cortex-connector.git\ncd cortex-connector\nnpm install\nCORTEX_ACCESS_TOKEN=<your token> node server/index.js\n```\n\nOr with Docker:\n\n```bash\ndocker build -t cortex-connector .\ndocker run -i --rm -e CORTEX_ACCESS_TOKEN=<your token> cortex-connector\n```\n\n| Variable | Meaning |\n|---|---|\n| `CORTEX_ACCESS_TOKEN` | Bearer token for the hosted connector (the OAuth access token issued at sign-in, for example the one `mcp-remote` stores). Optional: without it the server only lists tools. |\n| `CORTEX_MCP_URL` | Remote endpoint, defaults to the hosted Cortex connector. |\n\n`npm test` runs a self test (initialize, tools/list, one call without token).\n\n## The seven tools\n\n| Tool | Type | Description |\n|---|---|---|\n| **Search memories** | read | Semantic search across your memories |\n| **Fetch memory** | read | Full detail of one memory, with its sources |\n| **Recall & synthesize** | read | Narrative synthesis of what Cortex knows about a topic |\n| **Show conflicts** | read | Tracked contradictions between memories |\n| **Save memory** | write | Stores a memory — after a quality check against duplicates and redundancy |\n| **Write status** | read | Outcome of a queued save: quality-gate verdict and any open conflicts |\n| **Forget memory** | destructive | Deletes a memory — always with explicit confirmation |\n\n## Data & privacy\n\nMemories are personal: each user only sees their own. They are exportable and deletable on request, and the infrastructure operates in the European Union. Details: [Privacy Policy](https://skynetlab-cortex.com/privacy).\n\nTo disconnect, simply remove the connector from your client's settings; your memories remain in your Cortex account until you request their deletion.\n\n## Is this open source?\n\nThe Cortex memory engine is a **hosted, patent-pending service** — its source code is not published. This repository contains the public documentation, the client-side configuration examples and a small stdio bridge (`server/`) for connecting to it. Everything in this repository is released under the [MIT License](LICENSE).\n\n## Support\n\n📧 info@skynetlab.net · [Contact page](https://skynetlab-cortex.com/contatti)\n\n---\n\n*Cortex is a [SKYNETLAB](https://skynetlab.net/) product — independent research lab & deeptech studio, Bergamo, Italy. Founder: Filippo Pilotta ([ORCID 0009-0000-5002-4199](https://orcid.org/0009-0000-5002-4199)).*\n\n*Some content and product outputs are generated or assisted by artificial intelligence, pursuant to Art. 50 of Reg. (EU) 2024/1689. [AI transparency](https://skynetlab-cortex.com/trasparenza-ai).*\n",
  "bytes": 6845,
  "sha": "707e959f6f24f1892aafd12929b6a84cdb7b06c0aa203a007170e317214bca5f",
  "repo_slug": "filippopilo/cortex-connector",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_filippopilo_cortex_f073510d/readme"
}