One interface.
Every frontier model.

Meridian is model-agnostic by design. It doesn't make one model feel magical — it lets you choose, compare, combine, and route across every major AI system.

Choose, compare,
and combine

Every model is available through a single, consistent interface. @mention to route, or let the Brain pick the best one for your task.

Claude Opus & Sonnet
GPT-4o & o1
Gemini 2.0
Grok
NVIDIA NIM
Qwen
MiniMax
ByteDance Seedream
Xiaomi
OpenRouter
Custom API keys
+ More coming

Three ways to
use models

01

@mention routing

Type @claude or @gpt-4o in any message to send your prompt directly to that model. Mention multiple models to get parallel responses.

02

Brain auto-routing

Let Meridian's Brain classify your task and automatically select the best model based on capability, cost, and your history.

03

Everyone workflow

Send a prompt to all whitelisted models. They respond, build on each other, and debate until the strongest answer emerges.

04

Custom API keys

Bring your own keys for any supported provider. Use your own quota, get direct access, and keep full control over billing.

Every model,
one workspace

Start free with 10+ models. Upgrade to Pro for full frontier access and Brain routing.