OpenAI Responses API Now Ingests Excel, Word, and PowerPoint Directly

Following the global debut of the GPT-5.3-Codex programming model, OpenAI has concurrently expanded the repertoire of supported file formats within its API platform. This enhancement facilitates the seamless ingestion of a broader spectrum of document types during API invocations.

With this update, the OpenAI Responses API’s proficiency in handling professional documentation has ascended from mere utility to native fluidity. Developers are no longer burdened with the arduous task of manually converting documents into PDF format or flattening spreadsheets into CSV strings.

• Rich Document Formats: Support for DOC, DOCX, RTF, and ODT ensures the preservation of intricate paragraph structures, tables, headings, lists, and metadata such as headers and footers.

• Comprehensive Spreadsheets: XLS, XLSX, and CSV files are now automatically parsed—covering the initial 1,000 rows per sheet—while generating model-assisted summaries that maintain column headers and original formatting.

• Presentation Software: For PPT and PPTX, the model meticulously extracts text, titles, and bulleted lists on a per-slide basis.

By simply attaching these files as inputs within the Responses API call, developers empower the model to interpret structured content directly. OpenAI has pledged a continuous expansion of this library to encompass even more ubiquitous formats.

• Enterprise Automation: Human Resources may now ingest resumes, finance departments can process balance sheets, and legal teams can evaluate contracts directly for automated extraction, comparison, and report generation.

• RAG Augmentation: Retrieval-Augmented Generation knowledge bases are no longer confined to sterile PDF and text files; the entire Microsoft Office ecosystem now enjoys native compatibility.

• Efficiency and Fiscal Prudence: Organizations can now eschew third-party parsing libraries and conserve token expenditures previously squandered on rudimentary format conversions.