Toolang

A language and runtime for agents and humans.

The language expresses agent loops with fine-grained control using a small subset of natural language — readable by humans and agents, precise enough for the runtime to execute.

Getting started

Toolang runs on macOS and Linux with Python 3.11+. Windows is not supported yet.

Install Toolang with uv or pip; the main command is toolang, with too as a shorter alias:

uv tool install toolang  # or: pip install toolang
too --version            # same as: toolang --version

Models

Toolang bundles a catalog of recent models from models.dev, filtered by both knowledge > 2025-00 and last_updated > 2026-00. Set your provider's API key to get started; no additional model configuration is required. Local models are discovered from running Ollama and llama.cpp servers.

too providers --all  # List all providers and their required environment variables
too models           # List ready models

Download updated or full model catalogs from openhat-ai/models.

Shared agents

Use a shared agent by its URL or GitHub shorthand, with no agent configuration required:

too https://toolang.ai/lex.too chat  # Practice English with Lex
too briceyan/dev chat                # Write and review code with Dev

Toolang downloads the .too source and its declared dependencies and runs the agent locally. This lets you share an agent as a file, without separate installation or configuration instructions.

Local agents

Create an agent from scratch, or clone a shared one as a starting point:

too new NAME
too clone https://toolang.ai/lex.too NAME
too clone briceyan/dev NAME

Once created, the agent is ready for a conversation:

too NAME chat

Use the CLI help to configure the agent and add tasks or recurring chores:

too --help

Scripts

A .too file can also run as a script. Here is hello world in Toolang:

cat > hello-world.too <<'EOF'
agic():
  Say hello to the world.
EOF

too hello-world.too

To start from a template, use too init to create aide.too and toolang.toml in a directory. Rename or edit it as needed, then call it from Makefiles, CI jobs, or other scripts:

too init DIR
too DIR/aide.too --help

The generated script provides six named runnables for issues, fixes, pull request reviews, project overviews, recent updates, and localization. Select one by name and pass its arguments on the command line:

too DIR/aide.too whats_for
too DIR/aide.too whats_new                     # Updates from the past week
too DIR/aide.too whats_new since=v0.2.0
too DIR/aide.too review -- "https://github.com/OWNER/REPO/pull/123"
too DIR/aide.too update_i18n locale=zh-CN

Language

The tree-sitter-toolang repository provides the grammar and parser packages for Python, JavaScript, and Rust. See the syntax reference for the full language syntax.

See examples for runnable Toolang programs:

Common commands

Use these commands to run and manage agents. Add --help to any command for its usage and options.

too new <agent>                       # Create a local agent
too clone <ref> <agent>               # Clone an agent
too <agent> chat                      # Chat with an agent
too [agent] shell                     # Open a shell in the Toolang root or agent home
too serve <ref>                       # Run an agent in the foreground
too start <agent>                     # Start an agent in the background
too stop <agent>                      # Stop a running agent

too init <dir>                        # Create aide.too and toolang.toml
too [run] <file.too> [runnable]       # Execute a .too file; run is optional

too caps                              # List available caps
too tools                             # List available tools
too models                            # List available models
too providers                         # List available model providers

too --help                            # Show common commands
too more                              # Show additional commands

Links

• Website: toolang.ai
• Docs: toolang.ai/docs
• GitHub: github.com/openhat-ai/toolang