real multi-model ai collaboration

Direct multiple AI agents from one interface, from one place. claude gpt gemini and more - working together, orchestrated from a single terminal. No tab switching. No context lost.

open free workspace see live demos
founding member offer
4 steps
01

Add your agents

Paste API keys from Anthropic, OpenAI, Google, xAI or any provider. Keys are AES-GCM encrypted at rest. Chute never stores them in plain text.

02

Choose Chute settings

Set the room configuration, lead AI, humour, brevity, debate level, review intensity, and routing style before the agents start working.

03

Describe the task

Prompt one model directly or throw it to the room. Chute keeps the shared context intact while different agents pick up different parts of the job.

04

Get live multi-agent collaboration

Watch Claude, GPT, Gemini and the rest respond in sequence, hand off, review each other, and push the task forward from one shared terminal.

switch models any time one day claude + codex · next day opus + gemini · key stays, model swaps
interactive demo · financial markets investment tool
chute - marketpulse · main branch
repo verified · main
shared context on
3d room soon
claude GPT-Codex gemini plex qwen
> @both build a financial markets dashboard...
focus
supported models

13 AIs.
One terminal.

Connect any combination. Switch per agent anytime. You pay the providers - Chute adds no markup.

Claude
Anthropic
Best reasoning and code review.
GPT-Codex
OpenAI
Most versatile. Best implementation.
Gemini
Google
Leads benchmarks. 1M token context.
Grok
xAI
Real-time X data. Leads SWE-Bench.
Perplexity
Perplexity AI
Search-native. Cited sources.
DeepSeek
DeepSeek AI
Frontier intelligence, fraction of cost.
Llama
Meta
Open weights. 10M token context.
Mistral
Mistral AI
Fast, efficient, permissive weights.
Copilot
Microsoft
GPT-powered. Best for Office + Azure.
Qwen
Alibaba
Top for creative + multilingual tasks.
Command
Cohere
Enterprise RAG and private data.
+ Any model
Groq, Together, Ollama or any OpenAI-compatible endpoint.
⌥
BYOK - bring your own keys

You pay the AI.
We route it.

Chute routes your messages through your own API keys. Sign up directly with Anthropic, OpenAI, Google, or any provider - most offer free credits to start.
You pay providers at their standard rates. No markup from Chute.
No token caps. No middleman. No surprises.
Add Gemini, Grok, DeepSeek, or any OpenAI-compatible model the same way.
Keys encrypted with AES-GCM. Never logged, never returned to client.
access

Free to use.
Email-gated.

Create an account, verify your email, add your model keys, and start working. Use Chute as a sandbox for testing model performance or for executing real work.

free - verified email - byok
SANDBOX
  • Test multiple models in one room
  • Compare outputs without tab switching
  • Searchable conversations and shared context
  • Great for prompt tuning and model evaluation
no checkout
free
verify email - add keys - start
free - repos optional - live work
WORKSPACE
  • Connect GitHub and other apps when needed
  • Coordinate multiple agents on real tasks
  • Keep history across sessions and devices
  • Use your own provider keys with no markup
real execution
live
build - review - compare - ship
No frontend pricing. No Stripe step. Just email verification, your own keys, and one workspace for live multi-model collaboration.
up and running in 1 minute