Recall — Memory Agent for Students
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Recall — Memory Agent for Studentsrecall latest observations from biology lab"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
🧠 Recall — Memory Agent for Students
Amrita University MCP Hackathon 2026
Track: Education & Research
Team Members: [Insert Names & Roll Numbers]
Deployed URL: [Insert Nitrostack URL]
Demo Video: [Insert Video Link]
Problem & Solution
Recall is an MCP (Model Context Protocol) server that acts as a persistent memory agent for students. It captures narrated observations during labs, lectures, or fieldwork, stores them as structured, searchable memory, and lets the student — or an AI study assistant — recall them later across sessions.
Example: A chemistry student narrates "this is the titration setup for experiment 3, the flask needs to sit for 10 minutes" during a lab. Later, they (or their AI assistant) can recall that observation, get a study summary, or generate a lab report draft — all from the MCP server.
Related MCP server: clova-speech-lecture-mcp
🏗 Architecture
Built with the NitroStack MCP framework using a NestJS-style modular architecture:
recall/
├── src/
│ ├── index.ts # Bootstrap entry point
│ ├── app.module.ts # Root @McpApp module
│ └── modules/
│ ├── storage/
│ │ ├── storage.module.ts # Global storage module
│ │ └── database.service.ts # SQLite wrapper (sql.js)
│ └── recall/
│ ├── recall.module.ts # Feature module
│ ├── recall.tools.ts # 5 MCP tools
│ ├── recall.resources.ts # Session graph resource
│ ├── recall.prompts.ts # 2 prompt templates
│ ├── observation.service.ts # Business logic service
│ └── types.ts # TypeScript interfaces
├── data/ # SQLite DB (auto-created)
├── package.json
├── tsconfig.json
├── .env / .env.example
└── README.md🔌 MCP Primitives
Tools (5)
Tool | Description |
| Store a new observation with label, note, context, and session ID |
| Search observations by keyword, optionally scoped to a session |
| Edit an existing observation's note |
| Soft-delete an observation |
| List all sessions with counts and date ranges |
Resources (1)
URI | Description |
| Read-only structured dump of all observations in a session with inferred relations |
Prompts (2)
Prompt | Description |
| Generates a concise study summary from session observations |
| Turns session observations into a structured lab report skeleton |
🚀 Getting Started
Prerequisites
Node.js ≥ 20.18.1 (use nvm to manage versions)
npm ≥ 9
tsx installed globally:
npm i tsx -g
Install & Run
# 1. Install dependencies
npm install
# 2. Start the dev server
npx @nitrostack/cli devThe server starts in stdio mode, ready for any MCP client (NitroStudio, Claude Desktop, etc.) to connect.
Environment Variables
Copy .env.example to .env and adjust if needed:
DB_PATH=./data/recall.dbThe SQLite database and data/ directory are created automatically on first run.
📋 Example Usage Flow
1. Capture observations during a lab
→ capture_observation({
label: "titration setup",
note: "Flask with 25ml NaOH, add phenolphthalein indicator, 3 drops",
context: "experiment-3",
session_id: "chem-lab-2026-07-18"
})
← { id: "abc-123", timestamp: "2026-07-18T10:30:00Z" }
→ capture_observation({
label: "titration endpoint",
note: "Color changed from pink to colorless at 22.4ml HCl",
context: "experiment-3",
session_id: "chem-lab-2026-07-18"
})
← { id: "def-456", timestamp: "2026-07-18T10:45:00Z" }2. Recall observations later
→ recall({ query: "titration", session_id: "chem-lab-2026-07-18" })
← { count: 2, results: [ ... ] }3. Browse session graph
→ read resource: recall://sessions/chem-lab-2026-07-18/graph
← { observations: [...], relations: [...], metadata: { ... } }4. Generate a study summary
→ use prompt: study_summary({ session_id: "chem-lab-2026-07-18", topic: "titration" })
← [Formatted prompt with observations embedded, asking AI to produce a study summary]5. Draft a lab report
→ use prompt: lab_report_draft({ session_id: "chem-lab-2026-07-18" })
← [Formatted prompt asking AI to produce Objective → Method → Observations → Conclusion]🏆 Hackathon Rubric Alignment
Criterion | How Recall Addresses It |
MCP Primitives | All 3 implemented: Tools (5), Resources (1), Prompts (2) |
Education & Research | Memory agent designed for students in labs, lectures, fieldwork |
Code Quality | TypeScript strict mode, NitroStack decorators, DI, modular architecture |
Architecture | Clean separation: storage → service → tools/resources/prompts |
Runnable Locally |
|
Storage | SQLite via sql.js (WASM-based, auto-schema, swappable) |
🔮 Future Enhancements (if time allows)
Semantic similarity search (embeddings + vector store)
Camera/mic streaming for real-time observation capture
Canvas UI for visual browsing of observation graphs
Notion/Slack/Gmail integrations for exporting summaries
Graph DB backend (Neo4j/SurrealDB) for richer relation queries
📄 License
MIT
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceEnables transcription of audio files using Google Gemini AI and creates a searchable knowledge base, allowing users to ask natural language questions about content in meetings, podcasts, lectures, and other audio recordings.
- FlicenseNot gradedqualityDmaintenanceEnables speech-to-text transcription and summarization of lecture audio using Naver CLOVA APIs. Provides MCP tools for short and long audio processing with summarization.
- AlicenseNot gradedqualityDmaintenanceTurns any source material into a guided learning experience with learning maps, notes, four-phase study loop, spaced repetition flashcards, and grounded Q\&A.1MIT
- AlicenseNot gradedqualityDmaintenanceA local evidence recorder that captures screenshots, recordings, notes, and summaries during long-running agent tasks, enabling reviewable task records with timelines and artifacts.4MIT
Related MCP Connectors
Voice-led, FSRS-scheduled flashcards from YouTube, PDFs, web, or text. Auto-graded quizzes.
Search recordings, summarize meetings, create clips, and automate workflows from your AI assistant.
Search 4M+ podcasts & YouTube, transcribe any episode, search transcripts, generate AI lessons.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/sree-24066/amrita-mcp-recall'
If you have feedback or need assistance with the MCP directory API, please join our Discord server