docs: add OrcaRouter as an OpenAI-compatible provider example

[lovejones2914-spec]

Summary

Adala is an Autonomous DAta (L)abeling Agent framework — agents that acquire data-processing skills through iterative learning against a ground-truth environment, running on top of a litellm-backed OpenAIChatRuntime. Adala already documents how to use OpenAI-compatible gateways through OpenAIChatRuntime(provider="Custom", base_url=...), with OpenRouter shown as the current example. This PR adds OrcaRouter as a documented first-class provider, mirroring the existing OpenRouter example so users can point their skills at the OrcaRouter gateway without treating it as an anonymous custom base URL.

Changes

• Add an "OrcaRouter" example to the README Quickstart, directly after the existing OpenRouter section.
• Uses the same runtime wiring Adala already documents for OpenAI-compatible gateways: OpenAIChatRuntime(base_url="https://api.orcarouter.ai/v1", model="orcarouter/auto", api_key=os.getenv("ORCAROUTER_API_KEY"), provider="Custom").
• Documents the ORCAROUTER_API_KEY environment variable and the orcarouter/auto model, which routes each request to a live model.

Verification

• The new example is a full copy of the OpenRouter quickstart block, adjusted only for OrcaRouter's base URL, model, and env var, and it compiles cleanly.
• README fence balance validated (36 fences, balanced).
• Live test (L3): with ORCAROUTER_API_KEY from the environment, a real chat completion through the documented path (OpenAI(api_key=..., base_url="https://api.orcarouter.ai/v1"), model orcarouter/auto) returned HTTP 200 (routed to a live model, reply "ok"). The key was never printed or committed.

OrcaRouter is an OpenAI-compatible AI gateway built for both models and agents. Like OpenRouter, it exposes a provider/model namespace across many models — but it also combines adaptive routing, automatic failover, zero-markup inference, observability, guardrails, and agent-tool governance behind the same endpoint. Adding orcarouter as a first-class provider means this project's users can use that stack directly, without treating OrcaRouter as an anonymous custom base URL. It also runs gateway-level, zero-trust security for AI agents on the same endpoint — screening every prompt/response and governing every tool call on a default-deny basis, with no application code changes.

I'm an engineer on the OrcaRouter team. Questions or a different shape you'd prefer are welcome — Discord: discord.gg/YEubt8enRA · X: https://x.com/OrcaRouter