Quickstart
1. Install The Beta
Download the current macOS beta from the Document.Bot website:
https://document.bot/
Windows support is in alpha and should be treated as experimental.
2. Choose A Workspace
On first launch, pick a folder that contains the documents you want to work with. A good first workspace is small enough to understand but realistic enough to test the product:
- a project folder with PDFs and Word files
- a folder of manuals or policies
- a research folder with papers and notes
- a technical documentation folder with related references
Avoid starting with your entire home directory. The product works best when the workspace has a clear purpose.
3. Connect An AI Provider
Open Settings and connect the model provider you want to use. Depending on your build and configuration, Document.Bot can support online API providers, ChatGPT login, local models, or customer-hosted inference.
Use the provider that matches your data policy. For sensitive work, confirm whether your selected model sends content to an external service before using it on real documents.
4. Let The Workspace Index
Document.Bot indexes your files so search and AI workflows can use them. Indexing time depends on folder size, file formats, and your machine.
Start with a focused folder, wait for indexing to complete, then test a concrete query such as:
Find every place this requirement is mentioned and show me the source files.
5. Inspect Sources Before Acting
Use Document.Bot to open the original source behind an answer. For important work, treat retrieved text as a discovery layer and the original document as the source of truth.
The safest workflow is:
- Search the workspace.
- Open the source documents.
- Ask for a summary or draft.
- Review the evidence and any proposed changes.
- Accept, reject, or edit the output yourself.