Show HN: Use your OpenAI subscription to create audiobooks
Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.
This wires a document-to-audiobook pipeline into Codex via a Streamable HTTP MCP server, offloading TTS billing to your own OpenAI or Google keys rather than a separate SaaS fee. Practically, it's another MCP endpoint you can register today—just expect rate-limit and "high demand" failures on free Gemini keys, so enable billing if you want it to run reliably in any automated flow.
PDF-to-MP3 is packaged as an MCP server, so users can invoke audiobook generation directly from Codex while billing speech synthesis to their own OpenAI or Google account. For production agent builders, the notable pattern is “bring your own model billing” via MCP: it lowers the vendor’s COGS but pushes reliability, rate limits, and key-management issues onto the user’s provider account.