Redis agent memory explorer

A meeting-memory app that demonstrates Redis Agent Memory capabilities -- working memory, long-term memory, suggestions, and a conversational chatbot -- powered by the Redis Agent Memory (RAM) cloud service and Redis Context Surfaces for structured data retrieval.

Documentation

Detailed developer documentation lives in docs/dev-plans/:

Document	Covers
architecture.md	Monorepo layout, data flows, packages, env vars, cloud RAM behavior
backend.md	Routes, handlers, services, config, API contract
frontend.md	Components, state management, hooks, styling, dataset-driven UI
cau-ram.md	Cloud RAM wrapper API, operations, types, token budgeting
chatbot.md	LangGraph agent, tools, system prompt, CopilotKit integration
context-surfaces-integration.md	Context Surfaces setup, MCP tools, data flow, source attribution
suggestions.md	Real-time AI copilot pipeline, topic detection, stores

Prerequisites

• Node.js >= 18
• OpenAI API key for the chatbot, suggestion agents, and memory summarization
• Redis Agent Memory cloud credentials:
  • RAM_ENDPOINT -- REST API endpoint
  • RAM_API_KEY -- API key
  • RAM_STORE_ID -- store identifier
  • REDIS_URL -- Redis protocol URL for the same cloud instance
• Context Surfaces credentials (optional, enables structured data queries):
  • CTX_ADMIN_KEY -- Admin API key from Redis Cloud
  • CTX_SURFACE_ID -- Reuse an existing surface (auto-created on first run if blank)
  • MCP_AGENT_KEY -- Reuse an existing agent key (auto-created on first run if blank)

Quick start (Docker)

The fastest way to get the backend + frontend and LangGraph running in a single command.

1. Configure environment

cp .env.example .env

Edit .env and fill in the required values:

OPENAI_API_KEY=sk-your-actual-key-here

REDIS_URL=redis://your-cloud-redis-url:6379

RAM_ENDPOINT=https://your-ram-endpoint.redis.io
RAM_API_KEY=your-ram-api-key
RAM_STORE_ID=your-store-id

# Context Surfaces (optional -- enables structured data queries in chatbot)
CTX_ADMIN_KEY=ak_your-admin-key

2. Start all services

docker compose up --build

This spins up two containers:

Service	Description	Port
demo-app	Backend API + frontend (static build)	3001
demo-langgraph	LangGraph chatbot agent dev server	2024

Open http://localhost:3001 once all services are healthy.

Stop and clean up

docker compose down

---

Local development

If you prefer running services directly on your machine (e.g. for hot-reload).

1. Install dependencies and build packages

npm run setup

This runs npm install (resolves all workspaces) and then builds the shared packages/ libraries that the backend depends on.

2. Configure environment

cp backend/.env.example backend/.env

Edit backend/.env and fill in your credentials:

Variable	Description	Default
OPENAI_API_KEY	OpenAI API key	--
REDIS_URL	Redis protocol URL (cloud instance)	--
RAM_ENDPOINT	Redis Agent Memory cloud REST endpoint	--
RAM_API_KEY	Redis Agent Memory cloud API key	--
RAM_STORE_ID	Redis Agent Memory cloud store ID	--
CTX_ADMIN_KEY	Context Surfaces admin API key (optional)	--
CTX_ADMIN_API_URL	Context Surfaces admin REST endpoint	https://cloud.redis.io/context-surfaces
CTX_MCP_URL	Context Surfaces MCP endpoint	https://gcp-us-east4.context-surfaces.redis.io/mcp
CTX_SURFACE_ID	Reuse existing surface (auto-created if blank)	--
MCP_AGENT_KEY	Reuse existing agent key (auto-created if blank)	--
MEETING_MEMORY_PORT	Backend API port	3001
MEETING_MEMORY_ACTIVE_DATASET	Active dataset folder name	wealth-advisor
MEETING_MEMORY_MODEL_NAME	OpenAI model for suggestions and summarization	gpt-4o-mini
MEETING_MEMORY_CHATBOT_MODEL	OpenAI model for the chatbot agent	gpt-4o-mini
LANGGRAPH_DEPLOYMENT_URL	LangGraph local dev server URL	http://localhost:2024
LANGSMITH_API_KEY	LangSmith API key (optional)	--

3. Start everything

npm run dev

This single command builds the shared packages, then starts all three services concurrently with color-coded, labeled output:

Label	Service	Port
langgraph	LangGraph CLI dev server	2024
api	Express backend (hot-reload)	3001
frontend	Next.js dev server	3000

Open http://localhost:3000 once all services are up.

You can also start services individually in separate terminals if you prefer:

npm run dev:langgraph   # LangGraph agent graph
npm run dev:api         # Express API server
npm run dev:frontend    # Next.js frontend

Scripts reference

All scripts are run from the repo root.

Script	Description
npm run setup	Install all deps + build shared packages (run once)
npm run dev	Build packages + start all 3 services in one terminal

Individual services

Script	Description
npm run dev:frontend	Start Next.js dev server (port 3000)
npm run dev:api	Start Express backend with hot-reload (port 3001)
npm run dev:langgraph	Start LangGraph CLI dev server (port 2024)

Build and maintenance

Script	Description
npm run build:packages	Build only the shared packages/ libraries
npm run build	Build all workspaces
npm run test	Run tests across all workspaces
npm run clean	Clean build artifacts across all workspaces