# Blue Synergy Cortex

External cognitive memory engine backed by PostgreSQL and a knowledge graph. Keeps architectural requirements stable across agent sessions without drift.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 8
- Author: novotapeter-bs
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/novotapeter-bs/bs_cortex
- Updated: 2026-10-05
- HTML: https://affpapa.org/en/nevibecode/mcp/blue-synergy-cortex

## What it does

The server structures agent memory into distinct streams instead of an unorganized vector dump. A user stream registers conversational habits, while an authoritative knowledge stream locks architectural specifications into an ontology. Storing data inside transactional PostgreSQL and a domain graph prevents critical technical parameters from decaying during context compression.

During a standard interaction, the agent calls `cortex_recall` to fetch validated graph relations in under 65ms before writing code. Once the response completes, `cortex_ingest_turn` pipes the exchange into an asynchronous ingestion queue. For CLI environments like Claude Code, setting `compact: true` strips profile preambles to conserve context budget.

Local deployment is not supported. Without an active `ctx_live_...` API key, the hosted engine rejects queries, requiring constant network access to `cortex.bluesynergy.io`.

## Requirements
- Remote server: `https://cortex.bluesynergy.io/v1/mcp/sse`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport sse cortex https://cortex.bluesynergy.io/v1/mcp/sse
```

### Codex

**Codex**
- Codex connects MCP servers over stdio and streamable HTTP only, and this server speaks SSE

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "blue-synergy-cortex": {
            "url": "https://cortex.bluesynergy.io/v1/mcp/sse"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "blue-synergy-cortex": {
            "type": "sse",
            "url": "https://cortex.bluesynergy.io/v1/mcp/sse"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t sse blue-synergy-cortex https://cortex.bluesynergy.io/v1/mcp/sse
```

## When to use it
- Retain exact project specifications across separate terminal sessions
- Stop architectural requirements from drifting during prolonged tasks
- Retrieve domain graph relationships with latency under 65ms
- Separate user conversation tendencies from rigid system invariants

## Example prompts
- Call `cortex_recall` with `compact: true` to inspect database rules before drafting migrations
- Register a fresh microservice scope by running `cortex_create_project`
- Submit completed conversation turns for background ingestion using `cortex_ingest_turn`

## Similar
- [Penfield Memory](https://affpapa.org/en/nevibecode/mcp/penfield-memory.md) — MCP server · ★ 6
- [memory-continue](https://affpapa.org/en/nevibecode/skills/memory-continue.md) — skill · 1300 installs
- [pulse8-ai-cortex-knowledge-vault](https://affpapa.org/en/nevibecode/mcp/pulse8-ai-cortex-knowledge-vault.md) — MCP server · ★ 15
- [Memento](https://affpapa.org/en/nevibecode/mcp/scrypster-memento.md) — MCP server · ★ 9
- [robrain](https://affpapa.org/en/nevibecode/mcp/robrain.md) — MCP server · ★ 81
- [cortex](https://affpapa.org/en/nevibecode/mcp/cortex.md) — MCP server · ★ 18

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md