Skip to content

🎯 Quick Reference Guide

← Back to Home

Find the right AI tools for your use case quickly.

🔍 Find Tools by Use Case

"I want to build a chatbot"

  1. Choose a model: Qwen 2.5 or Llama 4
  2. Add memory: mem0
  3. Deploy: Ollama (local) or vLLM (production)
  4. Monitor: Langfuse

"I want to build a RAG application"

  1. Vector DB: Qdrant or Chroma
  2. Framework: LlamaIndex
  3. Embeddings: Use Transformers from HuggingFace
  4. Monitor: Helicone

"I want an AI coding assistant"

  1. IDE Extension: Continue or Cline
  2. CLI Tool: Aider or Plandex
  3. Skills: Browse SkillsMP.com for 66,500+ skills

"I need to fine-tune a model"

  1. Framework: Unsloth (2-5x faster)
  2. Dataset Tool: Argilla
  3. Training: PyTorch Lightning
  4. Track: Weights & Biases

"I want to build AI agents"

  1. Framework: LangChain or AutoGen
  2. Tools: MCP Servers
  3. Memory: mem0
  4. Monitor: LangSmith

"I need to evaluate my LLM"

  1. Framework: lm-evaluation-harness
  2. LLM-specific: DeepEval
  3. Monitor: Langfuse

💻 Find Tools by Hardware

"I have a laptop (8-16GB RAM)"

  • Model: Llama 3.2 8B, Gemma 4 (quantized)
  • Serving: Ollama
  • Vector DB: Chroma (embedded mode)

"I have a gaming PC (RTX 4090, 24GB VRAM)"

  • Model: Gemma 4 26B, Qwen 2.5 32B
  • Serving: vLLM or Ollama
  • Vector DB: Qdrant

"I have cloud/server access"

  • Model: Any (DeepSeek V3, Qwen 2.5 235B)
  • Serving: vLLM, TGI
  • Vector DB: Milvus, Weaviate

🎯 Find Tools by Experience Level

Beginner

Just starting? Go here: 1. Install Ollama - Easiest way to run models 2. Try Continue for coding help 3. Use Chroma for vector search

Intermediate

Have some experience? Try: 1. LangChain for building agents 2. Qdrant for better performance 3. Langfuse for monitoring

Advanced

Production-ready stack: 1. vLLM for serving 2. Milvus for scale 3. Ray for distributed computing 4. Langfuse + OpenLIT

🏢 Find Tools by Company Stage

Solo Developer / Prototype

  • Model: Ollama + Gemma 4
  • Framework: LangChain
  • Vector DB: Chroma
  • Monitor: Helicone (free tier)

Startup (< 50 users)

  • Model: Cloud API or self-hosted Qwen 2.5
  • Framework: LangChain or LlamaIndex
  • Vector DB: Qdrant
  • Monitor: Langfuse
  • Deploy: BentoML

Scale-up (50-10K users)

  • Model: vLLM serving Qwen 2.5 or DeepSeek V3
  • Vector DB: Qdrant or Weaviate
  • Monitor: Langfuse + LangSmith
  • Deploy: Kubernetes + vLLM
  • Workflow: Prefect

Enterprise (10K+ users)

  • Model: Multi-region vLLM deployment
  • Vector DB: Milvus (distributed)
  • Monitor: Full observability stack
  • Deploy: Ray + Kubeflow
  • Workflow: Prefect or Airflow

📊 Decision Matrix

Vector Database Choice

If you need... Choose...
Fastest setup Chroma
Best filtering Qdrant
Massive scale (billions of vectors) Milvus
Hybrid search Weaviate
SQL integration pgvector

LLM Observability

If you need... Choose...
Full-featured, free tier Langfuse
Fastest setup (15 min) Helicone
Deep LangChain integration LangSmith
OpenTelemetry standard OpenLIT

Deployment Platform

If you need... Choose...
Highest throughput vLLM
Easiest local setup Ollama
Production-grade SDK BentoML
HuggingFace ecosystem Text Generation Inference

🚀 Quick Start Paths

Path 1: Build a ChatGPT Clone (Weekend Project)

Day 1:
- Install Ollama
- Pull Qwen 2.5 model
- Set up basic chat interface

Day 2:
- Add Langfuse for monitoring
- Deploy with BentoML

Path 2: Build RAG Application (1 Week)

Week Plan:
Day 1-2: Set up Chroma + embed documents
Day 3-4: Integrate LlamaIndex
Day 5: Add chat interface
Day 6: Set up monitoring (Langfuse)
Day 7: Deploy and test

Path 3: Production AI System (1 Month)

Week 1: Infrastructure (vLLM + Qdrant)
Week 2: Application logic (LangChain)
Week 3: Monitoring & evaluation
Week 4: Testing & deployment (Kubernetes)

🔗 External Resources

← Back to Home