OrcaRouter as an AI Provider

Author: Tanishq-1Created Sep 11, 2026Updated Sep 11, 2026

OrcaRouter (https://www.orcarouter.ai/) is a multi-provider API gateway — one OpenAI-compatible API that routes to upstream providers (OpenAI, Anthropic, Google Gemini, DeepSeek, xAI Grok, Alibaba Qwen, Moonshot Kimi, MiniMax, Z.ai, Kling, ByteDance Seedance) at provider cost, with no per-token markup. It offers free access to some models through API keys, which fits FCC's free-provider support:

  • Free tier: models priced $0 under -free ids — currently deepseek/deepseek-v4-flash-free and deepseek/deepseek-v4-pro-free — plus the orcarouter/free router alias that auto-picks among the free models per request.
  • OpenAI-compatible Chat Completions at https://api.orcarouter.ai/v1 — streaming, tool calling, structured outputs, vision, and a unified reasoning_effort syntax (low/medium/high) translated per upstream.
  • API-key auth: Authorization: Bearer sk-orca-..., keys minted at https://www.orcarouter.ai/console.
  • Model discovery: GET /v1/models (OpenAI list shape, provider-prefixed ids like openai/gpt-4o-mini, anthropic/claude-sonnet-4.6).

Proposed wiring as a standard catalog provider: provider id orcarouter, env ORCAROUTER_API_KEY (+ optional ORCAROUTER_BASE_URL, ORCAROUTER_PROXY), default base https://api.orcarouter.ai/v1.

Docs:

Notes: free-tier rate limits are per-workspace fixed windows (RPM/day) tiered by lifetime spend, with a per-request prompt cap on the lowest tier; 429s return code free_rate_limited and use Retry-After to distinguish retryable windows from non-retryable prompt caps. GET /v1/models requires the key, so it can also serve as a credential-validation probe.

Source: Alishahryar1/free-claude-code