All use cases
Hermes Agent
Run Hermes Agent on AnyRouter
Install the AnyRouter provider plugin, add one key, and Hermes lists only the models that can call tools.
What this is for
Hermes Agent runs long tool loops, and a single upstream model means you eat that provider's outages and rate limits. Picking a model that cannot call tools breaks the loop.
Steps, then copy
- 1.Install the plugin with hermes plugins install anyrouter, then check it with hermes plugins list.
- 2.Add the key to ~/.hermes/.env as ANYROUTER_API_KEY=sk-.... Get one at anyrouter.dev; the same key also works against the OpenAI-compatible endpoint anyrouter.dev/api/v1.
- 3.Pick a model: run hermes model and choose anyrouter, or launch with hermes --provider anyrouter -m anyrouter/hermes. anyrouter/hermes auto-routes across the top tool-calling models with fallback, a good default for agentic work.
Install the plugin
bash
hermes plugins install anyrouter
hermes plugins list~/.hermes/.env
bash
ANYROUTER_API_KEY=sk-...Run Hermes on AnyRouter
bash
hermes --provider anyrouter -m anyrouter/hermesOptional settings
bash
# Point Hermes at a proxy or self-hosted relay
ANYROUTER_BASE_URL=https://relay.example.com/api/v1
# Credit Hermes traffic in the public app rankings
ANYROUTER_APP_ATTRIBUTION=1Ready to try it
The guide has the full walkthrough. Sign up if you still need a key.
Links
Next playbooks
- Run the Claude Agent SDK on any modelquery() already reads ANTHROPIC_BASE_URL and ANTHROPIC_API_KEY. Point both at AnyRouter — no SDK fork.
- Run every coding agent through one URLClaude Code, Codex, OpenCode, Cursor, and other OpenAI-compatible tools share 271+ models, one balance, and one request log.
- Give agents MCP tools without per-client sprawlUnderstand what an MCP server is, then connect agents to AnyRouter instead of pasting a new config into every app.