[aiapienthusiast]

Description

Add Cheaper Inference as an OpenAI-compatible text provider.

• use https://api.cheaperinference.com/v1
• use bearer authentication from CHEAPERINFERENCE_API_KEY
• support chat, streaming, and structured object generation
• reject all other operations with an accurate provider error
• add the provider to the README table with an official documentation link

Hardening

• use an explicit :chat and :object operation allowlist
• add a structured object request test
• test all public unsupported operation classes
• test the exact authorization header with an explicit test key
• merge the current main branch before validation

Testing

• focused provider tests: 12 passed
• full test run: 5,428 passed, 11 skipped, and 240 excluded; the release fixture had only a worktree dependency-path failure and passed after the normal local deps path was present
• mix quality: passed
• pre-push format hook: passed

Related work

• feat(cheaperinference): add provider catalog llmdb#370

Checklist

• Code follows the project style
• Documentation is updated
• Tests cover the new behavior
• Commits use Conventional Commits
• CHANGELOG.md is not changed

[mikehostetler]

Thanks for adding Cheaper Inference. Before this can merge, please open a matching PR in agentjido/llmdb that adds the Cheaper Inference provider and model catalog entries, then link that PR here. ReqLLM can use fallback model IDs without catalog data, but the LLMDB entries are needed for verified model metadata, pricing, token limits, and capabilities.

[mikehostetler]

Matching LLMDB catalog PR: https://github.com/agentjido/llmdb/pull/370\n\nIt adds the Cheaper Inference runtime contract and the gpt-5.4-mini and claude-sonnet-5 catalog records with gateway-specific limits and prices.