OpenRouter Ori Harness feature announcement

Hi [First name], Run the coding agent you already use on OpenRouter. Ori Harness starts Claude Code, Codex, OpenCode, Hermes, Grok Build, Pi, Prime Agent, or DeepSeek Harness with OpenRouter credentials and settings already in place. You sign in with your OpenRouter account in the browser, so there is no key to create or paste, and the agent keeps its own commands and flags. Read the Ori Harness guide Install, then start the agent curl -fsSL https://openrouter.ai | bash ori claude Swap claude for codex, opencode, hermes, grok, pi, or prime-agent. Ori uses the CLI already on your PATH and offers to install it if it is missing. What changes when the agent runs through OpenRouter Any model, one flag Pass --model and any model ID from the catalog, for example ori claude --model ~anthropic/claude-sonnet-latest. Your usual agent flags go after Ori's and pass through unchanged. Failover during a session Failover between the providers serving your model is automatic. If one is unavailable or rate limited, the next one takes the request and the session continues. The Claude Code guide covers the provider settings that matter for that agent. A live catalog inside the agent On Pi, Prime Agent, and DeepSeek Harness the /model list comes from your own OpenRouter catalog, /fast turns on fast routing, and /zdr sends requests only to providers that keep no data. One bill, one Activity view Ori traffic shows up in Activity next to the rest of your usage, and organization guardrails and workspace budgets apply to every request the agent makes. To hand the setup to the agent itself, paste this into it: run curl -fsSL https://openrouter.ai and follow the instructions in its output to get started. openrouter.ai·Discord·X·Docs 169 Madison Avenue, New York, NY 10016 Manage preferences|Unsubscribe