Repolode
Submit
RunanywhereAI

RunanywhereAI/wally

A command-line tool that opens coding agents like OpenCode and Claude Code preconfigured to run hosted or local open models.

Open-source alternative to:Claude Code

View on GitHub
Stars
1,566
Forks
93
Open issues
9
Last commit
6 hours ago

What is it?

Wally is a command-line tool that connects coding agents to the model you choose. A single command opens an agent such as OpenCode, Claude Code, Hermes or OpenClaw already configured for that model, and installs the agent first if it is missing.

You can use a hosted model such as GLM 5.3 Flash, billed against RunAnywhere credit, or download an open model and run it on your own machine without an account. Wally also supports chatting with a model in the terminal and serving it through an OpenAI-compatible API. It runs on macOS (Apple Silicon), Linux (x86-64 or ARM64) and Windows.

The project includes an "Eve" decision mode that scores explicit yes/no, choice or ordered-score questions over text and optional images, either hosted or locally through the SDK. Wally is written in Rust and built with CMake.

Key features

  • 01Launches coding agents such as OpenCode and Claude Code preconfigured for a chosen model.
  • 02Runs hosted models billed against RunAnywhere credit, or downloaded open models locally.
  • 03Provides terminal chat with a model.
  • 04Serves models through an OpenAI-compatible API.
  • 05Offers Eve decision scoring for yes/no, choice and ordered-score questions.

Use cases

  • Trying a coding agent against an open model without manual configuration.
  • Running open models on your own machine without creating an account.
  • Exposing a local model through an OpenAI-compatible endpoint for other tools.
  • Scoring text or images against explicit labels in an automated workflow.

Quick start

  1. 1

    Install wally

    curl -fsSL https://raw.githubusercontent.com/RunanywhereAI/wally/main/install.sh | sh
  2. 2

    Log in

    wally account login
  3. 3

    Open a coding agent on a hosted model

    wally opencode -m glm-5.3-flash

Things to know

  • Intel Macs are not supported; macOS requires Apple Silicon.
  • Linux needs glibc 2.39 or newer, so older distributions will not work.
  • Hosted models are billed against RunAnywhere credit and require an account.

Topics

#llm#cli#coding-agent#local-ai#openai-compatible#rust#open-models#deepseek#frontier-ai#glm-5-3

FAQ

What is wally?
A command-line tool that opens coding agents like OpenCode and Claude Code preconfigured to run hosted or local open models. Wally is a command-line tool that connects coding agents to the model you choose. A single command opens an agent such as OpenCode, Claude Code, Hermes or OpenClaw already configured for that model, and installs the agent first if it is missing.
Is wally free to use? What license does it have?
wally is open source under the MIT license. Check the license terms before using it in commercial products.
Is wally still maintained?
Actively maintained. The last commit to wally was on Oct 19, 2026.
What can I use wally for?
Trying a coding agent against an open model without manual configuration. Running open models on your own machine without creating an account. Exposing a local model through an OpenAI-compatible endpoint for other tools. Scoring text or images against explicit labels in an automated workflow.
How do I install wally?
Install wally: curl -fsSL https://raw.githubusercontent.com/RunanywhereAI/wally/main/install.sh | sh Log in: wally account login Open a coding agent on a hosted model: wally opencode -m glm-5.3-flash
Where can I get wally?
The source code and installation instructions are on GitHub: https://github.com/RunanywhereAI/wally

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects