memvault pro
Install, audit, and tune MemVault long-term memory for Claude Code, Cursor, and other MCP-compatible coding agents.
$15
Works with the AI Agents you already use
memvault pro
Example session with this skill installed
I need to set up MemVault for my current Claude Code project. I have Ollama running locally for embeddings. The binaries are in ~/tools/memvault.
- Read your context and instructions
- Compiled the memvault pro
I will configure MemVault as an MCP server for Claude Code using your local Ollama instance for embeddings.
Installation and Configuration
| Step | Action | Details |
|---|---|---|
| 1 | Directory Check | Verified ~/tools/memvault contains the server binary. |
| 2 | Data Path | Initialized ~/.local/share/memvault/data/current-project. |
| 3 | Embedding Config | Set provider to Ollama (http://localhost:11434/v1) using nomic-embed-text. |
| 4 | MCP Wiring | Executed setup_mcp.py to update .claude-code/config.json. |
| 5 | Tool Verification | Registered memory_write, memory_search, memory_get, and others. |
# Configuration command executed:
python scripts/setup_mcp.py --agent claude-code \
--memvault ~/tools/memvault \
--data-dir ~/.local/share/memvault/data/current-project \
--embed-url http://localhost:11434/v1 \
--embed-model nomic-embed-text \
--project .
Next steps
- Restart Claude Code to enable the
memory_*tools. - Seed the memory by providing 3-5 core project facts (e.g., "The build command is
npm run build:prod"). - Run
memory_searchto verify the injection and candidate table output.
Connects securely to your tools. The creator never sees your data.
About this skill
Managing persistent state across coding sessions is difficult. LLMs lose context, repeat mistakes, and forget project-specific architecture decisions. MemVault Pro solves this by integrating a disciplined, auditable long-term memory engine into your IDE or CLI agent. It transforms your agent from a stateless function into a collaborative partner that remembers your build commands, style preferences, and past refactors.
What it does
- Persistent Storage saves durable facts, architecture decisions, and project conventions across sessions.
- Auditable Recall explains why an agent knows a specific fact using retrieval IDs and evidence logs.
- Context Management injects relevant memories into the prompt based on semantic search and task relevance.
- Fact Lifecycle handles superseding stale information and irreversible deletion of forgotten facts.
- Multi-Agent Setup configures memory namespaces for Claude Code, Cursor, VS Code, Windsurf, and Claude Desktop.
How it works
- Installation involves downloading MemVault binaries, setting a stable data directory, and configuring embedding providers like Ollama.
- Wiring uses the provided setup script to merge MCP configurations into your specific agent's config file.
- Seeding initializes the memory with core project facts like deploy paths, code owners, and build scripts.
- Operation follows a loop of searching at task starts, writing durable learnings, and superseding contradictions.
Frameworks & tools
This skill manages the MemVault MCP server. It supports agents including Claude Code, Cursor, VS Code Copilot, Windsurf, and Claude Desktop. It integrates with Ollama for local embeddings or keyword-based search.
Why this beats prompting it yourself
Manual prompts for memory are noisy and lack versioning. This skill enforces a strict hygiene layer, preventing duplicate facts and providing a verifiable audit trail. It manages the technical overhead of namespace layouts and token budgets that simple "remember this" prompts ignore.
Use cases
- Cross-session continuity by recalling architectural constraints discussed days ago.
- Onboarding automation by seeding new workspace agents with established project style guides.
- Compliance auditing by proving specific data has been permanently deleted via the forget tool.
- Performance tuning by adjusting decay and retention settings when recall becomes noisy.
Known limitations
Keyword-only recall occurs if no embedding endpoint is provided. The data directory allows only one active process at a time due to ledger locking. Scripted config merges require plain JSON files.
Requires MemVault binaries. Ledger locks prevent concurrent CLI and MCP access to the same directory. Embedding provider required for semantic search. Setup script requires plain JSON config files.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 27 days ago
Requires