Claude Code · built into every workspace

The dev environment where
AI is the default

Developers open a workspace and Claude is already running — configured, context-aware and ready to code. No plugins. No API keys. No "let me set up AI real quick."

Get started →

Self-hosted on your infrastructure · your code never leaves your servers

coder@workspace — claude
coder@workspace:~/services/api$ claude

╭─────────────────────────────────────────╮
│ ✻ Welcome to Claude Code! │
│ Your AI pair programmer is ready. │
╰─────────────────────────────────────────╯

> add input validation to the registration endpoint

● I'll add validation to src/routes/auth.py...
● ✓ Updated src/routes/auth.py (+34 lines)
● ✓ Added tests in tests/test_auth.py (+28 lines)

All endpoints now validate email format, enforce
password length ≥ 8 and return RFC 7807 errors.

AI-native by design

Not bolted on. Built in.

Most teams spend hours wiring up AI tools per developer. CodeFlame does it once — at the infrastructure level — so every workspace is AI-ready from second one.

🤖
Claude Code, zero config
Every workspace launches with Claude Code pre-installed, authenticated and context-aware. Developers type claude and start working. No API keys, no plugins, no onboarding.
📄
Project context in CLAUDE.md
Each template ships a CLAUDE.md that tells Claude your stack, conventions and repo structure. Claude arrives already briefed — not asking "what framework do you use?"
🔌
MCP tools auto-configured
Jira, Confluence, GitHub and Bitbucket MCP servers are wired up from each developer's profile. Claude can read tickets, PRs and docs without leaving the terminal.
🖥️
VS Code in the browser
Full VS Code experience via code-server — extensions, themes, keybindings. Works on any device with a browser. SSH access for JetBrains Gateway is also built in.
📦
Custom Docker templates
Define workspace templates with custom images, pre-cloned repos and env vars. Write the Dockerfile in the UI and build it with one click. Build once, reuse for the whole team.
🏠
Self-hosted, fully owned
Runs on your own servers. Source code, credentials and Claude API keys never leave your infrastructure. No SaaS subscriptions, no usage caps, no surprise bills.

How it works

From zero to AI-assisted coding in under a minute

No VM setup, no "install the AI plugin", no sharing API keys with the team. Admins configure it once — developers just open a workspace and code.

Admin defines a template
Choose the Docker image, set repos to clone, env vars and — crucially — write a CLAUDE.md that tells Claude your architecture, conventions and team norms.
Developer starts a workspace
Pick a template, name the workspace, click Start. A container spins up with all tools pre-installed, repos cloned and Claude authenticated — in under 30 seconds.
Open the IDE — AI is already there
VS Code opens in the browser. A terminal tab is waiting with Claude running. No plugin to install, no key to paste — it's just there, briefed and ready.
Let AI handle the boilerplate
Write a feature, fix a bug, review a PR, create a Jira ticket — all in natural language from the terminal. Claude has your code, your tickets and your docs in context.

Integrations

Claude reads your tickets, PRs and docs too

MCP servers for Jira, Confluence, GitHub and Bitbucket connect automatically from each developer's profile. Claude has full context — not just your code, but the whole project.

🐙 GitHub
🪣 Bitbucket
📋 Jira
📝 Confluence
🔑 SSH
🐳 Docker
🤖 Claude Code

Ready?

Stop setting up AI. Start shipping with it.

Your whole team — one login away from an AI-powered workspace.
No per-developer setup, no API key chaos, no "it works on my machine."

Sign in →