Analysis updated 2026-05-18
Run an AI coding assistant in the terminal using your own API keys for Gemini, OpenAI, Anthropic, or DeepSeek.
Set up a team of AI models with different roles, such as architect, coder, and reviewer, that route tasks automatically.
Run the assistant headlessly inside a CI/CD pipeline to review pull requests or automate coding tasks.
| helis-d/waylou | adun-denton/chinvat | agbara286/saint-cms | |
|---|---|---|---|
| Stars | 2 | 2 | 2 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 3/5 | 4/5 | 2/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Node.js 20+ and an API key from at least one supported provider, or Ollama installed locally for offline models.
Waylou is an open source AI coding assistant that runs in your terminal, forked from Google's Gemini CLI. Where the original only connects to Gemini, Waylou lets you plug in several different AI providers through one interface, including Gemini, OpenAI, Anthropic, Ollama for local models, and DeepSeek. You bring your own API keys for whichever providers you want to use, and you can even switch models mid session. One of its bigger features is provider orchestration, which lets you set up a team of models with different roles, such as one acting as architect, one as coder, and one as reviewer, with tasks distributed between them using strategies like picking the best model for the job, round robin, or having several models vote. It also defines something called the ACE specification, short for Agentic Coding Environment, which standardizes how an AI agent is allowed to behave at three levels: basic file and shell operations, standard features like memory and sandboxing, and enterprise features like policy enforcement and audit logging. Waylou can run interactively in your terminal or headlessly, meaning it can be scripted into CI/CD pipelines and automated workflows without a human typing into it. For finding relevant code, it uses a context engine that combines vector embeddings with AST based code analysis. For safety, it can execute commands inside a sandbox using macOS Seatbelt, Docker, or Podman. Installing it requires Node.js 20 or newer and npm 9 or newer, and you can install it globally with npm or run it directly with npx. The project describes itself as under active development, with core functionality working but the provider layer still being built out, and the maintainers say that layer is where they most want help from contributors. It is released under the Apache License 2.0.
A terminal-based AI coding assistant forked from Gemini CLI that lets you bring your own keys and switch between multiple AI providers.
Mainly TypeScript. The stack also includes TypeScript, Node.js, CLI.
Use freely for any purpose, including commercial use, as long as you keep the copyright and license notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.