disclaimer

The funny numbers are estimates.

Last updated 18 August 2026.

• Not financial advice. Token counts, dollar amounts, and “99% saved” are worked examples of a typical agent loop (write script → install → fail → retry) versus a single tool call. Your model, prices, and prompt will differ.
• Latency depends on hardware, document size, images, and format. “Milliseconds” is the convert step, not the time an LLM spends writing Markdown.
• Compatibility. MCP and init-ai cover known agent config paths. We cannot inject a first-party tool into Claude, Codex, or anyone else.
• Quality. docgen is a compiler with themes. It is not a designer, a form-filler, or a replacement for tracked-changes Word work.
• Software. Provided as-is. See the repository LICENSE.
• Memes are satire, not a statement about any vendor’s product quality.

← back