RunAnywhere

What it does

RunAnywhere is the company behind Wally, a hosted open-model service, and a set of open-source on-device SDKs. Wally takes a chat request from your app, an OpenAI client or a coding agent and streams back a model reply through one OpenAI-compatible endpoint, or through the wally CLI, which starts a supported coding agent on a chosen model.

For on-device work, one C++ core sits behind SDKs for Swift, Kotlin, React Native, Flutter, TypeScript and C++, so the same capabilities (LLM, speech-to-text, text-to-speech and vision) run inside your own app on the user's hardware. Developers using Wally keep their existing OpenAI SDK code and change the base URL, key and model ID.

This suits developers and technical teams who want open models without managing serving infrastructure, or who want models running inside their apps. Hosted use requires a RunAnywhere account and prepaid credits; direct API calls need a Cloud key. Entitlements can differ between environments for the same model ID, and current per-token rates are published in the console.

What it can do

• Hosted open-model chat completions through an OpenAI-compatible endpoint
• Streaming replies as the model generates tokens
• One CLI command that starts OpenCode, Claude Code, Hermes or another supported agent on a chosen model
• Local or hosted model selection with the same wally commands
• Open-source on-device SDKs for Swift, Kotlin, React Native, Flutter, TypeScript and C++ over one C++ core
• On-device LLM, speech-to-text, text-to-speech and vision-language model calls

Works with

• OpenCode
• Claude Code
• Claude Desktop
• Codex
• Command Code
• Hermes
• OpenClaw
• DeepSeek Harness
• Prime Agent
• Dify

Before you choose

• Entitlement is per environment: a key that works in development can return 403 in production for the same model ID.
• Hosted commands use credits; local model runs stay on your machine, and Wally does not move requests between local and cloud on its own.
• The console playground keeps your recent conversations, unlike the default zero-retention production behaviour.

Listed under

• Coding
• Developer tools
• LLM inference
• Npu acceleration
• On-device AI
• Openai-compatible API
• Open-source SDK

Get started

Create an account, buy prepaid credits and create a Cloud key, or install the CLI and sign in.

1. Install the wally CLI with the install script for your platform, then run wally account login.
2. Create a Cloud key in the console for direct API calls and point your OpenAI client at https://inference.runanywhere.ai/v1.
3. Call GET /v1/models with your key and use a returned model ID rather than a hardcoded one.

Hosted usage is billed per million tokens against prepaid credits bought in the console. Check the console for current per-model rates and the pricing page for plan details.

We couldn’t find published prices for RunAnywhere. See their pricing page ↗

Sources for this overview

Product pages, documentation, and pricing checked 30 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Wally: frontier open models, one API key ↗runwally.com
• Open Source | RunAnywhere ↗runanywhere.ai
• Wally documentation - RunAnywhere Docs ↗docs.runanywhere.ai
• OpenAI-compatible API - RunAnywhere Docs ↗docs.runanywhere.ai
• RunAnywhere ↗www.runanywhere.ai

Company website

runanywhere.ai: product, pricing and policy pages

Hacker News

Launch posts and discussion

YC directory mirror

Batch and company name

Web search

Confirms the official site

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

manual_review

Last checked 30 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.