Package Exports

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (hilla) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

Hilla CLI

AI co-founder in your terminal. Plan on canvas at hilla.ai, execute in CLI.

  ▄▄▄        ▄▄▄▄     ▄▄▄▄▄
▄█████▄    ▄██████▄ ▄████████▄
████████▄▄▄█████████████████████
 ▀█████████████████████▀▀   ▀██▀
    ▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀

hilla.ai is the brain — you sketch your project there: goals, stages, features, tech stack. hilla CLI is the hands — it takes that plan, sets up a sandbox, and writes real code into it.

Chat and execution are bound to a pulled Hilla project. Load work with /pull before asking questions or running tasks.

Install

npm install -g hilla

Requires Node 18+.

Quick start

# 1. Launch the TUI
hilla

# 2. Connect your Hilla account (create a token at hilla.ai/settings/tokens)
/login

# 3. Pull a project from your canvas
/pull            # opens the project picker
/pull 7          # direct pull still works

# 4. Let the agent start building it
/run             # ask mode — you approve each file write / shell op

# Plain text is chat only, even after a project is loaded
What should I build first?

# Or launch directly in autonomous mode:
hilla --all-in

# Slash commands can also be launched from your shell:
hilla pull 7
hilla run

Every project gets its own sandbox at ~/.hilla/sandboxes/<project-id>/. The agent only reads and writes inside this directory. Metadata (canvas context, history, plan) lives in ~/.hilla/sandboxes/<project-id>/.hilla/.

Agent modes

Execution starts only from /run. Plain text messages are chat, and can answer questions about the loaded project without editing files.

Ask mode (default): every write_file, edit_file, and bash call prompts you for y/n.
Autonomous mode (hilla --all-in): file reads/writes can run without prompting, but shell commands still require approval because they are real local processes. Ctrl+C cancels at any time.

Commands

Command	Description
`/login`	Connect your Hilla account
`/logout`	Sign out, remove local credentials
`/whoami`	Show current user
`/projects [n\|keyword]`	Browse projects or pull a selected project
`/pull [n\|keyword]`	Open the project picker or pull directly into sandbox
`/run [instruction]`	Start agent loop on current project. Bare `/run` picks the next step.
`/ask <question>`	One-off chat question (no tools)
`/cd`	Show the current project's sandbox path
`/status`	Current project summary
`/tasks`	Show tasks of loaded project
`/model [list\|n\|name]`	Show or switch AI model
`/credits`	Credit balance and plan
`/history`	Last 20 commands
`/config`	Show current settings
`/exec <cmd>`	Run a safe read-only command inside the current sandbox
`/clear`	Clear the chat
`/quit`	Exit

Agent tools

Inside /run, the model has access to:

read_file(path) — read inside the sandbox
write_file(path, content) — create/overwrite (asks approval)
edit_file(path, search, replace) — surgical diff edits
list_dir(path) — list files
bash(command, reason) — run shell inside sandbox with a scrubbed environment (always asks approval)
task_done(summary) — mark the task complete
ask_user(question) — pause and ask for clarification

Paths are relative to the project sandbox. Any attempt to read or write outside ~/.hilla/sandboxes/<id>/, including symlink escapes, is rejected. The backend verifies the loaded project and owns the execution tool schema; custom tools and generic no-project agent requests are rejected server-side.

Keyboard

Tab — switch view, or complete slash commands
↑ / ↓ — command history
Ctrl+S — toggle sidebar
Ctrl+L — clear the chat
Ctrl+C — cancel agent turn / approval prompt

AI models

/model list to see available models. Switch with /model <n|name>.

sonnet-4.6 — coding
opus-4.6 — heavy reasoning
kimi-k2.5 — writing / Turkish
gemini-3.1-pro-preview — planning
glm-5-turbo — fast coding
minimax-m2.7 — agentic
deepseek-r1 — research

All model labels route to the matching OpenRouter model ID through Hilla's backend and are metered server-side — your account token is the only credential.

Config

~/.hilla/config.json:

{
    "defaultModel": "sonnet-4.6"
}

~/.hilla/auth.json holds your PAT (chmod 600). ~/.hilla/config.json stores non-secret preferences only.

Environment overrides: HILLA_API_URL, HILLA_PUBLISHABLE_KEY.

Develop

git clone https://github.com/hillaai/hilla-cli.git
cd hilla-cli
npm install
npm run build
npm link
hilla

hilla