Toolang

Toolang is a programming language and runtime for defining and executing agent loops in natural language.

Toolang is a language and runtime for agents and humans that structures agentic work. It allows users to define agent loops as functions called agics, compose them for complex tasks, and execute them on a computer or in a sandbox. The system uses natural language that humans, large language models, and the runtime can read and write. Users can orchestrate agics in sequence, in parallel, or in loops, and extend agents with composable primitives like psyches, skills, services, and prompts. Toolang supports cloud models from providers such as Alibaba, Anthropic, DeepSeek, Google, OpenAI, OpenRouter, Vercel, and local models served by Ollama and llama.cpp.

Toolang at a glance

Best for: Software engineer
Runs on: macOS, Linux
Pricing: Free
Released: November 2026
Category: AI coding tool
Website: toolang.ai

Key features

Define agent loops as functions
Orchestrate agics in sequence or parallel
Extend agents with composable agent primitives
Use natural language plain-text files
Execute agents on computers or in sandboxes
Inspect active runs and execution history
Support cloud and local models

Use it for

Create your own agent
Run a shared agent
Run agent scripts in sandboxes
Monitor agent runs and execution history

Alternatives to Toolang

pstack — Improves AI coding agent outcomes with rigorous workflows.
LangChain — Compose exactly the agent your use case needs from model, tools, prompt, and middleware.
OpenViking — Unifies agent memory, knowledge RAG, and skills into a single directory structure.
HarnessRouter — One API to manage agent runtimes, sandboxes, and execution lifecycles.
AGENTS.md — It separates agent-specific instructions from human-focused README files.
Alook — Connects existing local coding agents to team communication channels.
Toolmaker Agent — It runs locally as a single binary and is entirely free to use.
AG-UI — Standardizes agent-user interactions and generative UI for applications.
MyAgent — It features built-in permission controls, long-term memory, and skill self-evolution.
Codeg — Run multiple AI coding agents together and manage them from mobile devices.
Langroid — Simplifies multi-agent orchestration and app development with intuitive abstractions.
Agentloop — Provides a clean agent loop with provider-agnostic LLM access, hooks, and safety budgets.

All Toolang alternatives