free-router
humantonylee/free-router
Terminal tool that pings free AI models and writes the best one into OpenCode, OpenClaw or Hermes Agent
free-router is a TypeScript command-line tool that discovers free AI models, measures their live latency and uptime, and configures a coding agent to use the one you pick. Hermes Agent is one of its three configuration targets.
What free-router does
free-router is a terminal interface that lists free models from NVIDIA NIM and OpenRouter and pings all of them in parallel every 2 seconds. Each row shows rank, a capability tier derived from SWE-bench score, provider, context window, rolling average latency, latest latency, session uptime and a short verdict. The default ranking puts availability first, then higher tier, then lower latency.
Pressing Enter on a model configures OpenCode, OpenClaw or Hermes Agent to use it, and a search bar filters the list. A first-run wizard asks for provider API keys and opens the provider sites in the browser; ESC skips a provider, and a keyless ping still shows latency. Keys are read from environment variables first, then from ~/.free-router.json. Running free-router --best prints the ID of the best responding model so scripts can use it.
Key features
- Parallel ping of every listed free model every 2 seconds with live latency, uptime and verdict
- Capability tier per model derived from SWE-bench score
- Enter in the list or the search bar configures OpenCode, OpenClaw or Hermes Agent
- First-run wizard for NVIDIA NIM and OpenRouter keys, plus a settings screen with live key tests
- free-router --best prints the best responding model ID for scripts
When to use it
- Pointing Hermes Agent at a free model that is actually responding right now
- Comparing latency and uptime across free NVIDIA NIM and OpenRouter models before choosing one
- Picking a model ID in a shell script with the non-interactive --best flag
Who it is for: Developers who run Hermes Agent, OpenCode or OpenClaw on free model tiers and want a quick way to find a working model.
How it fits with Hermes Agent
free-router supports Hermes Agent as one of three configuration targets next to OpenCode and OpenClaw, so it is a general model picker rather than a Hermes-only tool.
How to install free-router
These commands are copied from the project's README. Check the repository for the latest steps before you run them.
npx @bytonylee/free-router
npm i -g @bytonylee/free-router
bunx @bytonylee/free-router
bun install -g @bytonylee/free-routerRequirements: Free API keys for NVIDIA NIM (nvapi- prefix) and OpenRouter (sk-or- prefix) are optional; without a key the tool still shows ping latency.
FAQ
What is free-router?
free-router is a command-line tool that discovers free AI models from NVIDIA NIM and OpenRouter, pings them live, and configures OpenCode, OpenClaw or Hermes Agent to use the one you select.
Does free-router work with Hermes Agent?
Yes. Selecting a model and pressing Enter offers Hermes Agent as one of the configuration targets, alongside OpenCode and OpenClaw.
Is free-router free and open source?
Yes. The repository is licensed under Apache-2.0, and the models it lists are free-tier models from NVIDIA NIM and OpenRouter.
Similar models for Hermes Agent
All modelsAI middleware for model routing, RAG and OpenAI-compatible APIs, with Hermes Agent sync hooks
VasiHemanth TokenTelemetryLocal dashboard for token usage, cost and tool calls across coding agents, with a dedicated Hermes Agent view
soulduse AI Token MonitorMenu bar and system tray app that prices your AI coding tool usage from local logs, Hermes Agent included
y13sint FreeQwenApiLocal OpenAI-compatible proxy that exposes a Qwen Chat web account as an API for tools such as Hermes Agent
bitrouter BitRouterLocal Rust model router with an OpenAI-compatible API that learns routing policies for agent workflows
icoretech Codex PoolerSelf-hosted gateway that fronts Codex accounts with stable Pool API keys for agents and teams
Related guides: How to install Hermes Agent