Prism
The Mind Palace for AI Agents - HIPAA-hardened Cognitive Architecture with on-device LLM (prism-coder:7b), Hebbian learning, ACT-R spreading activation, adversarial evaluation, persistent memory, multi-agent Hivemind and visual dashboard. Zero API ke
What is Prism?
Prism is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to mind palace for ai agents - hipaa-hardened cognitive architecture with on-device llm (prism-coder:7b), hebbian learning, act-r spreading activation, adversarial evaluation, persistent memory, multi-ag...
The Mind Palace for AI Agents - HIPAA-hardened Cognitive Architecture with on-device LLM (prism-coder:7b), Hebbian learning, ACT-R spreading activation, adversarial evaluation, persistent memory, multi-agent Hivemind and visual dashboard. Zero API ke
This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- The Mind Palace for AI Agents - HIPAA-hardened Cognitive Arc
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx prism-mcpConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Prism
Prism MCP is a HIPAA-hardened cognitive memory architecture for AI agents that stores and retrieves session context, code knowledge, and long-term facts using a combination of local SQLite and optional cloud sync. It exposes 40+ MCP tools implementing episodic, semantic, and procedural memory lanes, Hebbian-style reinforcement, ACT-R spreading activation for relevance scoring, and drift detection to flag when an agent's work diverges from its original goals. Designed for local-first operation with zero API keys required, it optionally integrates on-device LLMs via Ollama (prism-coder models) and supports multi-agent Hivemind coordination.
Prerequisites
- Node.js 18+ or Bun runtime installed
- An MCP-compatible client such as Claude Desktop, Cursor, or Claude Code
- Ollama installed locally if you want to use the on-device prism-coder LLM models (optional)
- A Synalux API key if you want cloud sync instead of local-only storage (optional; local SQLite works by default)
Add Prism MCP to your client configuration
No separate installation step is needed — npx downloads and runs the package automatically. Add the following entry to your MCP client config file.
{
"mcpServers": {
"prism": {
"command": "npx",
"args": ["-y", "prism-mcp-server"]
}
}
}Choose a storage backend (optional)
By default Prism uses local SQLite with no configuration needed. To enable cloud sync, set PRISM_STORAGE and provide your Synalux key. To force local-only regardless of other settings, set PRISM_FORCE_LOCAL=true.
# For cloud sync:
export PRISM_STORAGE=synalux
export PRISM_SYNALUX_API_KEY=synalux_sk_...
# For explicit local-only:
export PRISM_FORCE_LOCAL=trueInstall an on-device LLM (optional)
If you want Prism's cognitive routing to use a local model instead of a remote API, install Ollama and pull one of the prism-coder models sized for your hardware.
# Minimal devices (1.1 GB):
ollama pull dcostenco/prism-coder:1b7
# Default router (8.4 GB, recommended):
ollama pull dcostenco/prism-coder:14bRestart your MCP client
Quit and reopen Claude Desktop or reload the MCP server in Cursor. Prism's 40+ tools (session_load_context, knowledge_search, session_detect_drift, etc.) should appear in the tool list.
Load context at the start of a session
Begin every coding session by asking your assistant to load prior context. Prism will surface relevant memories, prior decisions, and session state from its knowledge store.
Prism Examples
Client configuration
Full claude_desktop_config.json entry for Prism MCP with local SQLite storage (no API keys needed).
{
"mcpServers": {
"prism": {
"command": "npx",
"args": ["-y", "prism-mcp-server"],
"env": {
"PRISM_STORAGE": "auto"
}
}
}
}Prompts to try
Example prompts that exercise Prism's memory, drift detection, and knowledge ingestion tools.
- "Load my context for the 'checkout-refactor' project and summarize what we worked on last session"
- "Store the fact that we decided to use Postgres instead of MySQL for the orders table"
- "Search my knowledge base for everything related to authentication middleware decisions"
- "Detect if my current work has drifted from the original goal I set at the start of this session"
- "Ingest the files in /src/services so I can query them semantically later"Troubleshooting Prism
npx fails to find or run prism-mcp-server
Ensure Node.js 18+ is installed and `npx --version` works. If you are behind a corporate proxy, set the NPM_CONFIG_PROXY environment variable. You can also install globally first with `npm install -g prism-mcp-server` and then use `command: prism-mcp-server` in the config.
Ollama model not found when using local LLM routing
Run `ollama list` to confirm the model was pulled successfully. Ensure the LOCAL_LLM_URL environment variable points to your Ollama instance (default is http://localhost:11434). Ollama must be running before starting Prism.
Cloud sync fails with Synalux authentication error
Verify your PRISM_SYNALUX_API_KEY starts with `synalux_sk_` and has not expired. Set PRISM_STORAGE=local as a fallback to continue working locally while troubleshooting the cloud connection.
Frequently Asked Questions about Prism
What is Prism?
Prism is a Model Context Protocol (MCP) server that mind palace for ai agents - hipaa-hardened cognitive architecture with on-device llm (prism-coder:7b), hebbian learning, act-r spreading activation, adversarial evaluation, persistent memory, multi-agent hivemind and visual dashboard. zero api ke It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Prism?
Follow the installation instructions on the Prism GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Prism?
Prism works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Prism free to use?
Yes, Prism is open source and available under the AGPL 3.0 license. You can use it freely in both personal and commercial projects.
Prism Alternatives — Similar Knowledge & Memory Servers
Looking for alternatives to Prism? Here are other popular knowledge & memory servers you can use with Claude, Cursor, and VS Code.
MemPalace
★ 52.6kA local AI memory system that stores all conversations verbatim and organizes them into navigable structures. It provides 19 MCP tools for AI assistants to search and retrieve past decisions, debugging sessions, and architecture debates automatically
Kratos
★ 25.7k🏛️ Memory System for AI Coding Tools - Never explain your codebase again. MCP server with perfect project isolation, 95.8% context accuracy, and the Four Pillars Framework.
Context Mode
★ 15.4kAn MCP server that preserves LLM context by intercepting large data outputs and returning only concise summaries or relevant sections. It enables efficient sandboxed code execution, file processing, and documentation indexing across multiple programm
Memu
★ 13.7kMemory for 24/7 proactive agents like OpenClaw.
MemOS
★ 9.3kMemOS (Memory Operating System) is a memory management operating system designed for AI applications. Its goal is: to enable your AI system to have long-term memory like a human, not only remembering what users have said but also actively invoking, u
Everos
★ 5.4kBuild, evaluate, and integrate long-term memory for self-evolving agents.
Browse More Knowledge & Memory MCP Servers
Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Prism in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Prism?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.