Multi-model workspace

Roxy AI

1M contextBest fit for huge repositories, long specs, and image inputs.

Markdown-first chat + OpenAI-compatible API

Better chat UI, real code rendering, file previews, and a drop-in API.

Roxy keeps the original glassy theme, but upgrades it into a proper coding workspace: markdown replies, syntax-highlighted code blocks, image/file attachments, and an owner-managed API key system.

Markdown + code blocks

Assistant replies render like a real coding chat, including fenced language blocks.

10MB attachments

Drop screenshots, docs, or text files into the chat and image parts render inline.

OpenAI-style API

Use `/api/chat/completions` or `/api/v1/chat/completions` with Bearer keys.

Models

Long-context coding stack

GLM 4.5

Zhipu

128K context

Fast, roomy context for long edits and big pasted files.

Gemini 2.5 Pro

Google

1M context

Best fit for huge repositories, long specs, and image inputs.

Llama 3.3 70B

OpenAI-compatible endpoints

Base URL

/api

curl /api/chat/completions \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-2.5-pro",
    "messages": [{"role":"user","content":"Write a landing page hero."}],
    "stream": false
  }'

Also available at /api/v1/chat/completions and /api/v1/models for tooling that expects a versioned base path.

Read docs

API Keys

Owner key + client keys

Unlock with your owner key to create stored client keys. The owner key skips rate limits.