IdeaWell

Project ideas from Hacker News discussions.

The ChatGPT/Codex app bundles a full copy of LibreOffice

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1. Bloat and performance concerns
Many commenters criticize the size of the bundled LibreOffice headless runtime and its impact on disk usage, startup time, and overall system responsiveness.

  • “Bundling all of LibreOffice seems like a pretty huge dependency.” – pseudosavant
  • “libreoffice‑headless still shows as 429.7 MB.” – zamadatix
  • “I think it's overbloating my computer.” – winstonp
  • “2 gb is crazy town, these things could be around 100 mb.” – cpursley

2. Preference for lighter or native alternatives
Several users suggest that OpenAI could avoid the heavyweight dependency by using headless CLIs, native libraries (e.g., Apache POI, Pandoc, Rust), or relying on the user’s existing Office installation and its APIs/CLIs.

  • “Why not computer use or using the tools the user prefers instead? The primary office suites have APIs/CLIs now.” – verdverm
  • “That's where the CLIs come in for me, I'm not going to waste tokens on screenshots and input device manipulation.” – verdverm
  • “Apache POI” – simongg (mentioned as a lighter alternative for reading/writing Office files)
  • “I still don’t understand why with all their cash these companies dont redo their apps in Rust or native.” – cpursley
  • “I have a skill that instructs ChatGPT/Codex to verify the rendering using the Mac versions of Microsoft Office.” – telotortium

3. Licensing and distribution compliance
A notable thread discusses whether bundling LibreOffice respects its open‑source licenses (MPL/GPL), whether the binary is truly bundled or downloaded on demand, and the need to provide proper attribution.

  • “Curiously, I didn't find any reference in the Open source licences section of the codex app. Is this a MPL 2.0 violation?” – vb‑8448
  • “Despite the title, I'm not sure it's actually bundled. Some of the skills plugins will download it on the first run.” – zamadatix
  • “I finally did find the license in …” – zamadatix (showing the license file location)
  • “The license itself may be present in the cache folder.” – TZubiri
  • Discussion of needing to “provide attribution (in a way that doesn't bother the user)” to stay compliant.

🚀 Project Ideas

Generating project ideas…

OfficeBridge Daemon

Summary

  • A lightweight local service that exposes a simple HTTP/JSON API for creating, editing, and converting Microsoft Office and LibreOffice documents by delegating to the user's installed Office suite (COM on Windows, AppleScript on macOS) or a headless LibreOffice fallback.
  • Eliminates the need to bundle large runtimes with AI agent applications while providing reliable, high‑fidelity document manipulation.

Details

Key Value
Target Audience Developers building LLM‑powered coding agents, automation scripts, or AI‑assisted productivity tools that need to read/write .docx, .xlsx, .pptx files.
Core Feature Unified API endpoints (/convert, /extract, /update) that route calls to the best available local Office engine (native COM/AppleScript when present, otherwise headless LibreOffice) and return results as JSON or file streams.
Tech Stack Rust (for cross‑platform daemon), Tokio + Hyper for HTTP server, windows-sys/objc bindings for Office automation, optional libreoffice headless subprocess; can be packaged as a single binary or Docker image.
Difficulty Medium
Monetization Hobby

Notes

  • HN users complained about Codex bundling a 429 MB headless LibreOffice just to avoid UI interruptions ("Computer use is slow and can take control of the application focus"). OfficeBridge lets agents use the user's own Office install, avoiding focus theft and token‑wasting screenshots.
  • By detecting existing Office/LibreOffice installations, the daemon addresses the concern about unnecessary bloat ("Bundling all of LibreOffice seems like a pretty huge dependency"), while still providing a guaranteed fallback for headless conversion on Linux or headless‑only environments.

DocAgent CLI

Summary

  • A single cross‑platform command‑line tool that wraps Microsoft 365 CLI, Google Workspace CLI, and libreoffice --headless to perform common document operations (convert formats, extract text/data, update cells, generate previews) and outputs structured JSON.
  • Designed for AI agents to invoke via subprocess with minimal overhead and no UI focus stealing.

Details

Key Value
Target Audience AI agent developers, DevOps engineers, and power users who need scriptable, reliable office‑file handling without launching heavy GUIs.
Core Feature Sub‑commands like docagent convert --input file.docx --output file.pdf, docagent extract --input file.xlsx --sheet 1 --format json, docagent update --input file.pptx --replace "TODO" "DONE"; each command detects the best available backend (MS 365 CLI, Google Workspace CLI, or LibreOffice) and returns a JSON payload with status, artifacts, and any logs.
Tech Stack Go (for fast single‑binary distribution) with plugins for each CLI; uses exec.Command to call the underlying tools, parses their output, and normalizes to JSON. Optional WASM plugin for environments without native binaries.
Difficulty Low
Monetization Hobby

Notes

  • Commenters highlighted the usefulness of existing CLIs ("Why not computer use or using the tools the user prefers instead? The primary office suites have APIs/CLIs now"). DocAgent consolidates those CLIs into a predictable interface, saving agents from guessing which tool to call.
  • By returning JSON, the tool solves the frustration of needing to parse screen output or deal with focus stealing ("Suddenly it decides to open the app and try to do something + screenshot it"). Agents can remain headless while still leveraging powerful, native office functionality.

DocWasm

Summary

  • A compact WebAssembly library (≈300 KB) that provides core parsing, editing, and export capabilities for Office Open XML formats (.docx, .xlsx, .pptx) using pure Rust code compiled to WASM.
  • Exposes language‑agnostic bindings (Python, Node.js, Rust) so LLMs can manipulate documents in‑memory without spawning external processes or bundling heavy desktop apps.

Details

Key Value
Target Audience LLM‑agent frameworks, AI‑powered SaaS products, and extension developers who need fast, sandboxed document manipulation inside worker threads or browser‑based agents.
Core Feature In‑memory API: load a .docx/.xlsx/.pptx blob, traverse its XML structure, modify text/styles/tables/cells, and export back to the same or another format (PDF via optional tiny WASM PDF emitter). All operations run inside the WASM sandbox with deterministic latency.
Tech Stack Rust → wasm‑bindgen → WASM; Python package via pyodide or wasm-pack; Node.js wrapper; optional tiny WASM‑based PDF writer (e.g., pdfium subset). Builds to < 500 KB wasm file.
Difficulty High
Monetization Hobby

Notes

  • Several commenters wished for lighter alternatives to bundling LibreOffice ("There are already some very fast cross platform (wasm too!) libraries that can do what poppler does as well as parse and produce various office file formats"). DocWasm directly answers that call with a WASM‑native solution.
  • By avoiding external processes, the library sidesteps UI focus and token‑wasting concerns raised by users ("Computer use is slow and can take control of the application focus"), while still giving agents the power to read/write complex office files safely inside their sandbox.

Read Later