Coding plans billed monthly, with no rolling five-hour windows

typed is an AI coding CLI with a 1M context window on both hosted tiers and zero data retention on model inference for every hosted request. Your prompts and code are never used for training.

See plans, from $10 a month

No account needed to buy.

or run it free on your own machine via llama-server or Genie Server

Plans, per month

Starter $10 Pro $20 Max $100 Ultra $200

Every plan includes both hosted tiers on one budget, which resets each month. If you run out sooner, the next request returns an optional one-click top-up prompt.

With no typed.cloud key the CLI runs typed-local (free, on your machine). With a key on a paid plan it runs typed (hosted economy, billed to your plan). Name typed++ when you want the strongest tier.
Tier	Context	Use
typeddefault with a paid plan	1M	Economy: the lowest draw per token.
typed++strongest	1M	Hard tasks: `typed --model typed++`

Team plans start at $25 a seat, with hosted codebase indexing.

Free and local

typed-local runs entirely on your machine via llama-server or Genie Server: no account, no card, and your prompts never leave it. Setup guide

Install with one line

macOS, Linux, or Git Bash on Windows:

curl -fsSL https://app.typed.cloud/install.sh | bash

Windows, in PowerShell:

irm https://app.typed.cloud/install.ps1 | iex

Windows, inside WSL Ubuntu:

curl -fsSL https://app.typed.cloud/install.sh | bash

or from PowerShell, into your Ubuntu distro:

& ([scriptblock]::Create((irm https://app.typed.cloud/install.ps1))) -Wsl

With llama-server the free tier needs about 32 GB of RAM, since the model runs on your machine (requirements). With less, the same line still installs the launcher; hosted has no hardware requirement. On Windows the PowerShell line needs Git for Windows and offers to install it with scoop; the lines do not run in cmd.

With a plan, run typed login, then t in any project. Without one, run typed local, start the server it prints, then t.