Codex Slides: open-source AI slide studio powered by Codex. Prompt, repo to deck

Codex Slides is an open-source AI slide studio from the nexu-io/codex-slides repository on GitHub. It is powered by Codex and is positioned around a prompt-plus-repo-to-deck workflow: you describe the deck and point the tool at a codebase or related project material, then it produces slides. The project surfaced on Hacker News with 3 points and 0 comments at the time of the listing, under a Claude/Codex/Fable source thread.

Technically, the product is framed as an AI slide studio rather than a one-shot chat export. The core mechanic is generation from natural-language prompts combined with repository context, so deck structure, section titles, and technical talking points can be grounded in actual code and project layout instead of only in a free-form brief. That puts Codex in the role of the generation engine for layout and content assembly from repo-aware inputs.

For engineers and builders, the useful part is reducing the usual gap between a shipped codebase and a presentable deck. Internal design reviews, launch walkthroughs, architecture readouts, and investor or stakeholder updates often start from the same repo the team already maintains. A prompt-and-repo path can turn that material into a first draft of slides without rebuilding the narrative by hand from scratch.

In market terms, Codex Slides sits in the growing set of AI presentation tools, but with an open-source distribution model and an explicit repo-to-deck angle. Many competing products emphasize chat-to-slides or template fill-ins; this one markets the codebase as a first-class input. Early HN traction is light—3 points and no comments yet—so visibility is still limited relative to more established AI deck products.

Practical takeaway: if your workflow already lives in Git and you regularly need decks that match the system under discussion, clone nexu-io/codex-slides and test prompt-plus-repo generation against a real project you know well. Watch how faithfully it maps modules, APIs, and architecture into slide structure, and how much manual cleanup remains before a deck is shippable. Next signals worth tracking are community uptake on the GitHub repo and whether the HN thread gains deeper technical discussion beyond the initial listing.