[aiapienthusiast]

This adds a Cheaper Inference AI adapter.
Cheaper Inference is an OpenAI-compatible LLM gateway.
One API key gives access to models from several labs.
It uses the same pattern as #870 (Chutes) and #267 (LiteLLM).

• New package packages/ai/cheaperinference (@profullstack/sh1pt-ai-cheaperinference, adapter ID ai-cheaperinference).
• The code follows the current Chutes adapter: dry-run short-circuit, usage token mapping, clean base URL validation and redacted error excerpts.
• The API key comes from the CHEAPERINFERENCE_API_KEY vault secret.
• The chat URL is https://api.cheaperinference.com/v1/chat/completions.
• ai in packages/cli/src/adapter-registry.ts lists the new directory, so the registry test stays green.

Details

• Default model: gpt-5.4-mini.
  Other listed models: gpt-5.4, claude-sonnet-5, deepseek-v4-flash, glm-5.3.
  All ids come from the public model list (https://api.cheaperinference.com/public/models).
• baseUrl config is optional, like the other integrated adapters.
• src/index.test.ts mirrors the Chutes tests: smoke test, dry-run, request body and usage mapping, base URL normalization, unclean base URL rejection, redacted errors.
• README.md uses the same format as the generated module READMEs.
• pnpm-lock.yaml has the new workspace importer entry.

Test:

pnpm vitest run packages/ai/cheaperinference/src/index.test.ts packages/cli/src/adapter-registry.test.ts
pnpm --filter @profullstack/sh1pt-ai-cheaperinference typecheck

Docs: https://cheaperinference.com/docs