Verity AI Backends Compared - Groq vs OpenRouter vs Ollama

So Verity talks. That’s the whole point. But how it talks depends on which AI backend you wire it up to. Verity JE supports three options: Groq, OpenRouter, and Ollama. Two are cloud APIs, one runs entirely on your own hardware. Each has tradeoffs.

Here’s the quick version, then we’ll dig in.

The Three Options at a Glance

	Groq	OpenRouter	Ollama
Type	Cloud API	Cloud API (multi-model)	Local (your PC)
Speed	Blazing fast	Depends on model	Depends on your hardware
Quality	High	High (pick any model)	Varies by model size
Cost	Free tier available	Free tier available	Completely free
Privacy	Messages go to Groq servers	Messages go to OpenRouter	100% local, nothing leaves your PC
API key needed	Yes	Yes	No
Setup	Quick (get key, paste in config)	Quick (get key, paste in config)	Medium (install Ollama, pull model)
Best for	Fast responses, low-end PCs	Flexibility, model shopping	Privacy, offline play, no limits

Ollama — Run It Locally

Privacy first. Every word you say to Verity stays on your machine. Nothing hits the internet. If you care about that kind of thing — and you probably should, given Verity reads your inventory, health, coordinates — Ollama is the way.

Free forever. No API key, no usage limits, no rate caps. Download the model, run it, done.

Works offline. Once you’ve pulled the model, you don’t need internet at all. Play on a plane, play in a bunker, whatever.

Needs hardware. The AI runs on your PC. If you’re on a potato, responses will be slow. You want at least 8GB RAM, ideally 16GB. NVIDIA GPU helps a lot.

Setup takes a few extra steps. Install Ollama, pull a model, configure the mod. Not hard, but it’s more work than pasting an API key.

Full walkthrough: Ollama Setup Guide

Who it’s for: People with decent hardware who want full privacy and zero ongoing cost.

Groq — The Fast Cloud Option

Insanely fast. Groq’s LPU hardware cranks out responses quicker than you can read them. If Verity feels sluggish on other backends, Groq fixes that.

Great for low-end PCs. All the AI processing happens on Groq’s servers, so your laptop doesn’t break a sweat.

Free tier exists. Groq offers a generous free tier with rate limits. Enough for casual play.

Needs an API key. Sign up at console.groq.com, grab your key, paste it in the mod config. Easy, but it’s a step.

Your conversations leave your PC. Every message goes to Groq’s servers and comes back. They have privacy policies, but it’s not the same as running locally.

Getting Your Groq API Key

1. Go to console.groq.com and sign in (create an account if you don’t have one)
2. Click API Keys in the left menu
3. Click Create API Key
4. Name it anything — “Minecraft” works fine
5. Copy the key (it’s long, so use Ctrl+C)
6. Back in Minecraft, open the mod config and paste it into the Groq API key field

Who it’s for: Players on weaker hardware who want snappy responses and don’t mind cloud dependency.

OpenRouter — Pick Any Model

Note: OpenRouter is not listed on the official Modrinth page. The mod officially supports Groq and Ollama. OpenRouter is mentioned here based on community reports — verify compatibility before committing to it.

Model flexibility. OpenRouter isn’t tied to one provider — route to different models without changing anything on your end.

One API key for everything. Sign up once, use any model they offer.

Also cloud-based. Same hardware benefits as Groq — runs fine on any PC.

Needs an API key. Same deal — sign up at openrouter.ai.

Not officially confirmed. Check the Modrinth page or Discord before relying on this.

Getting Your OpenRouter API Key

1. Go to openrouter.ai and sign up
2. Click Get API Key → New Key
3. Name it (e.g., “Minecraft”) and click Create
4. Copy the key
5. Paste it into the mod config’s OpenRouter API key field

Finding models: After signing in, go to Models → Text. You can sort by pricing (low to high) or search “free” to find free models. Copy the model name and paste it into the mod config.

Who it’s for: Tinkerers who want to experiment with different AI models and don’t want to download 40GB each time.

Which One Should You Pick

Honestly? It comes down to two questions:

Got a decent GPU (NVIDIA, 4GB+ VRAM)? Go with Ollama. Free, private, unlimited. The setup is 10 minutes of work and then you never think about it again.

Running on a laptop or older PC? Go with Groq. Fast responses, free tier, minimal setup. You’ll need to grab an API key, but that’s a one-time thing.

You can always switch later — just change the AI provider setting in the mod config and you’re done.

Recommended Models for Groq/OpenRouter

If you’re using a cloud API, here are the models Varmite tested in the 6.0.0 setup video:

Model	Best For	Notes
GPT 120B	Smartest responses	Highest quality, but slower
Llama 3.1 8B Instant	Fastest responses	Optimized for speed
Llama 3.3 Versatile	Balanced	Varmite’s pick — good mix of speed and quality

For OpenRouter, you can also search “free” in their model browser to find models that won’t cost anything (though they still have rate limits).

Thinking Mode

The mod config has a Thinking Mode toggle. Varmite recommends turning it off — it makes responses faster. When thinking mode is on, the AI spends more time reasoning before responding, which adds latency without much benefit for casual gameplay.

How to Configure Each Backend

Specific configuration details depend on the mod version. Check your mod’s config file (supports TOML or JSON5 format) for the available AI provider settings.

• Ollama (local): Full step-by-step setup in our Ollama Setup Guide
• Groq (cloud): Refer to the SETUP section on the Modrinth project page
• OpenRouter: Not officially listed — check the Verity Discord for community guidance

Need help with the full installation? Check the Install Guide — it covers everything from Forge setup to AI backend configuration.

---

Still have questions?

---

Independent fan site - not affiliated with or endorsed by Mojang, Microsoft, ThatMob, CurseForge, or Modrinth.

By Verity Mod Hub Team · Last updated: July 2026