GLM-5.1 · Chain-of-Thought · Live

AI Chat with
Deep Reasoning

Powered by NVIDIA NIM & GLM-5.1. File uploads, multi-model support, and real chain-of-thought thinking — all in your browser.

⬡ Start Chatting → View Models
axecodi — chat
👤
Write a Python function to find all prime numbers up to n using the Sieve of Eratosthenes
⬡
Here's an efficient implementation:
def sieve(n):
  primes = [True] * (n+1)
  primes[0] = primes[1] = False
  for i in range(2, int(n**.5)+1):
    if primes[i]:
      for j in range(i*i,n+1,i):
        primes[j] = False
  return [i for i,v in enumerate(primes) if v]
Time complexity O(n log log n) — optimal for large n. ✓
👤
Explain what transformers are in ML
⬡
Transformers are neural network architectures based on self-attention — they process all tokens in parallel rather than sequentially, capturing long-range dependencies efficiently. Used in GPT, BERT, and virtually all modern LLMs...
✦ Features

Everything you need
in one chat interface

Built for developers and power users. No bloat, just the tools that matter.

🧠
Chain-of-Thought Reasoning
Watch GLM-5.1 think step-by-step. Collapsible reasoning block shows the model's internal logic before answering.
📎
File & Image Upload
Attach images, code files, CSVs, PDFs and more. Paste images directly from clipboard with Ctrl+V.
⚡
Real-time Streaming
Responses stream token-by-token for a fast, natural feel. Watch the answer build in real time.
🤖
11+ Models
Switch between GLM-5.1, Llama 3.3 70B, DeepSeek R1, Mistral, Gemma, Qwen, and more with one click.
💬
Multi-Conversation
Create and switch between multiple conversations. All history saved locally in your browser.
⌨️
Slash Commands
Type / to access commands: /model, /temp, /system, /export, /clear, /retry and more.
🔒
Secure API Proxy
API calls routed through Cloudflare Edge — no CORS errors, API key protected server-side.
📤
Export Chats
Download conversations as JSON, Markdown, or plain text. Keep your AI interactions.
🎨
Markdown + Syntax Highlight
Full markdown rendering with code syntax highlighting, tables, and one-click copy for code blocks.
🤖 Models

Top AI models,
one interface

All hosted on NVIDIA NIM infrastructure for low latency and high reliability.

Flagship
GLM-5.1
Best reasoning + thinking
Meta
Llama 3.3 70B
Powerful open-source LLM
DeepSeek
DeepSeek R1
Advanced reasoning model
Mistral
Mixtral 8x7B
Fast MoE architecture
Google
Gemma 2 9B
Lightweight & capable
+6 more models →
🚀 Deploy

Live in 3 minutes

Deploy your own instance on Cloudflare Pages for free.

1
Get API Key
Sign up at build.nvidia.com and get your free NVIDIA NIM API key.
build.nvidia.com → API Keys
2
Deploy to Cloudflare
Upload this project to Cloudflare Pages. Takes under 2 minutes.
pages.cloudflare.com → Upload
3
Set Env Variable
Add your API key as NVIDIA_API_KEY in Cloudflare Pages settings.
Settings → Env Variables
4
Go Live!
Your AI chat is live at yourname.pages.dev — share it with anyone!
yoursite.pages.dev ✓

Ready to start?

Jump into the chat — no setup needed if API key is already configured.

⬡ Launch Axecodi.ai.ai →