Deploy production-grade RAG pipelines in minutes, not weeks.
R8R is an end-to-end RAG workflow runtime that eliminates boilerplate and lets you build, run, and scale retrieval-augmented generation systems with a single API call.
Building RAG systems is painfully repetitive:
|
|
Drag, drop, and deploy RAG pipelines without code. Connect retrievers, rerankers, and LLMs visually.
- Redis for session context
- Qdrant for semantic memory (95.7% duplicate detection)
- PostgreSQL for full conversation history
- Automatic deduplication and consolidation
Run GPT-4, Claude, and Gemini in parallel. Get ensemble answers with automatic fallback and voting.
GPT-4 ──┐
Claude ──├──► Intelligent Ensemble ──► Optimized Answer
Gemini ──┘
Hypothetical Document Embeddings reduce hallucinations and improve context quality by 40%.
Build and deploy RAG systems via natural language chat. No dashboard required.
Track cost, latency, errors, and replay any query. Built-in debugging for production systems.
npm install r8r-clientpip install r8r-clientClone the repository and set up your local environment:
# Clone the repository
git clone https://github.com/Nitinref/R8R.AI.git
cd R8R.AI
# Install dependencies
npm install
# Start Qdrant (Vector Database)
docker run -p 6333:6333 qdrant/qdrant
# Start Redis (Caching)
docker run -p 6379:6379 redis:alpine
# Or use Docker Compose to start all services
docker-compose up -dCreate a .env file in the root directory:
# Database
DATABASE_URL="postgresql://neondb_owner:npg_8TjRiXy9GmYZ@ep-wild-bar-ad7o2d1l-pooler.c-2.us-east-1.aws.neon.tech/neondb?sslmode=require&channel_binding=require"
# Security
JWT_SECRET="your-super-secret-jwt-key-change-in-production"
JWT_EXPIRY="7d"
# LLM API Keys
OPENAI_API_KEY="sk-proj-your-key-here"
ANTHROPIC_API_KEY="sk-ant-your-key-here"
GOOGLE_API_KEY="AIzaSy-your-key-here"
MISTRAL_API_KEY="your-key-here"
# Vector Databases
PINECONE_API_KEY="pcsk_your-key-here"
PINECONE_ENVIRONMENT="gcp-starter"
PINECONE_INDEX_NAME="default"
WEAVIATE_URL="http://localhost:8080"
# Redis Configuration
REDIS_HOST="localhost"
REDIS_PORT="6379"
REDIS_PASSWORD="your_redis_password_here"
# Qdrant Configuration
QDRANT_HOST="localhost"
QDRANT_PORT="6333"
QDRANT_API_KEY="your_qdrant_api_key_here"
# Telegram Bot (Optional)
TELEGRAM_BOT_TOKEN="your-telegram-bot-token"
TELEGRAM_CHAT_ID="your-telegram-chat-id"
# Rate Limiting
RATE_LIMIT_WINDOW_MS="900000"
RATE_LIMIT_MAX_REQUESTS="100"
# Server Configuration
PORT="3001"
NODE_ENV="development"
NEXT_PUBLIC_API_URL="http://localhost:3001"Start the development server:
# Run database migrations
npm run db:migrate
# Start the backend
npm run dev:backend
# In another terminal, start the frontend
npm run dev:frontendYour R8R instance will be available at:
- Frontend: http://localhost:3000
- API: http://localhost:3001
- Qdrant Dashboard: http://localhost:6333/dashboard
import R8R from "r8r-client";
const r8r = new R8R("YOUR_API_KEY");
const result = await r8r.query(
"How does photosynthesis work?",
{
pipeline: "advanced",
memory: true,
providers: ["gpt-4", "claude"]
}
);
console.log(result.answer);
console.log(result.sources);
console.log(result.confidence);from r8r_client import R8R
r8r = R8R("YOUR_API_KEY")
result = r8r.query(
"How does photosynthesis work?",
pipeline="advanced",
memory=True,
providers=["gpt-4", "claude"]
)
print(result.answer)
print(result.sources)curl -X POST https://api.r8r.ai/v1/query \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"query": "Explain blockchain technology",
"pipeline": "standard",
"memory": true
}'Choose the right pipeline for your use case:
| Pipeline | Best For | Features | Response Time |
|---|---|---|---|
| standard | FAQs, chatbots, simple queries | Basic retrieval + single LLM | ~2s |
| advanced | Technical docs, complex Q&A | HyDE + reranking + ensemble | ~4s |
| research | Academic research, deep analysis | Multi-hop reasoning + citations | ~8s |
| custom | Anything | Visual builder, full control | Varies |
R8R's memory system works across three layers:
Fast in-memory storage for current conversation. Sub-10ms retrieval.
Vector embeddings with 95.7% duplicate detection accuracy. Finds relevant context from past conversations.
Complete conversation history, user preferences, and system metrics.
Benefits:
- ✅ Remember user preferences across sessions
- ✅ Eliminate redundant questions automatically
- ✅ Build context-aware applications
- ✅ Background consolidation (no blocking)
Query multiple LLMs simultaneously for faster, more reliable results.
Combine responses using weighted voting based on confidence scores.
If one provider fails, seamlessly fall back to alternatives.
Route queries to the most cost-effective provider for each task.
const result = await r8r.query("Question", {
providers: ["gpt-4", "claude", "gemini"],
strategy: "ensemble", // or "fastest" or "cheapest"
fallback: true
});Build RAG systems without leaving Telegram:
You: /create Build a customer support RAG using GPT-4 with memory
Bot: ✅ Created workflow "customer-support-rag"
🔑 API Key: r8r_abc123...
🌐 Endpoint: https://api.r8r.ai/v1/pipelines/cs-001
Test it:
curl -X POST https://api.r8r.ai/v1/pipelines/cs-001 \
-H "Authorization: Bearer r8r_abc123..." \
-d '{"query":"How do I reset my password?"}'
Available commands:
/create- Build a new pipeline/list- View all pipelines/stats- Get usage analytics/edit- Modify existing pipeline
- Query volume and response times
- Cost tracking per query and provider
- Error rates and failure patterns
- User satisfaction scores
Reproduce any query exactly as it ran in production for debugging.
Compare pipeline configurations with built-in experiment tracking.
Get notified when latency spikes, costs exceed thresholds, or errors occur.
Frontend
- Next.js 14 + TypeScript
- Tailwind CSS + shadcn/ui
- Canvas-based workflow editor
Backend
- Node.js + Express
- PostgreSQL + Prisma ORM
- Qdrant (vector database)
- Redis (caching & sessions)
Infrastructure
- Docker + Kubernetes
- AWS / GCP support
- Self-hosted option available
|
$0/month
|
$49/month
|
Custom pricing
|
All plans include built-in memory, analytics, and observability.
Build intelligent support bots that remember user issues and provide contextual help.
Let users query your docs naturally. Automatic updates when docs change.
Multi-hop reasoning across academic papers with citation tracking.
Connect Slack, Notion, Google Drive, and more. One search interface for everything.
Stop rebuilding the same RAG logic Focus on your product, not infrastructure.
Ship 10x faster From idea to production in hours, not weeks.
Scale with confidence Built-in monitoring, fallbacks, and cost controls.
Pay for value, not tokens Predictable pricing based on queries, not token usage.
No credit card required. Deploy your first RAG pipeline in under 5 minutes.
Built for developers who value their time. Deploy RAG systems that just work.