- Home
- Skills
- Agents & Orchestration
- memorius
Works with the AI Agents you already use
memorius
A self-contained vector memory vault for AI agents to store, search, and recall context across sessions.
Free
memorius
Example session with this skill installed
I finally fixed that race condition in the token refresh logic. It was a missing mutex lock in auth.ts. Store this for me.
- Read your context and instructions
- Compiled the memorius
Memory detected: Bug fix for token refresh race condition.
Store this? (y/n)
✅ Stored: "Fix: race condition in token refresh logic. Root cause: missing mutex lock in auth.ts. Solution: added mutex lock."
Shelf: bugs
Folder: authentication
Note: token-refresh-fix
Next steps
- Run
memorius search "token refresh"to verify retrieval. - Use
memorius diaryat the end of the session to document the full implementation. - Consider running
memorius consolidateif you have other notes on auth concurrency.
Connects securely to your tools. The creator never sees your data.
About this skill
Agents lose context across sessions, leading to repetitive questions and lost insights. Memorius provides a persistent, self-contained memory vault that allows any AI agent to store, search, and recall information using vector embeddings and session diaries.
What it does
- Vector search finds relevant memories based on semantic meaning rather than just keyword matches.
- Auto-capture triggers detect bug fixes, learning moments, decisions, and preferences to suggest storage.
- Context injection retrieves past work and project details to inform current agent responses.
- Session diaries generate summaries of accomplishments and decisions at the end of long work sessions.
- Information mining extracts structured memories from meeting transcripts, documents, and PDF files.
- Vault management includes tools for consolidating duplicates, checking facts, and organizing data into shelves and folders.
How it works
- Store memories via CLI or agent hooks by providing content and optional shelf/folder metadata.
- Search the vault using natural language queries to retrieve the most relevant past entries with confidence scores.
- Inject context into the current session by running the context command to feed previous findings back into the agent prompt.
- Consolidate the vault periodically to merge similar entries and maintain a clean, high-signal memory set.
Frameworks & tools
Built as a standalone CLI tool and MCP server. It uses local vector embeddings (all-MiniLM-L6-v2) for privacy and supports integrations with Obsidian, Claude Desktop, and common developer shells.
Why this beats prompting it yourself
Generic system prompts for "remembering" rely on the finite context window of a single session. Memorius offloads this to a persistent local database, allowing for multi-week project recall that doesn't bloat your token usage.
Use cases
- Bug tracking to remember root causes and fixes for recurring architectural issues.
- Developer preferences to persist coding styles, preferred libraries, and workflow patterns.
- Meeting synthesis to extract action items and decisions from long chat transcripts or logs.
- Project onboarding to quickly inject all previous context when starting a new task in an old repo.
Known limitations
Requires local storage access at ~/.memorius/data. Shell hooks require explicit user permission to modify shell configuration files. Large vaults (>100 memories) require periodic manual consolidation.
Requires local storage access. Shell hooks modify shell configs. Large vaults need periodic consolidation. No built-in encryption for stored memory 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
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 22 days ago
- Free with an Agensi account
Permissions
Need to install memorius pypi package