Choose a practical scenario
Start with a task you regularly repeat, rather than documenting everything at once. The guides below explain what to keep in Library, what to write in Context, how to use the result in everyday work, and how to check an update. Each includes three copyable practice pages, a linked relationship diagram, a completion check, and optional read-only AI prompts. You do not need to write software or connect an AI service.
Find your starting point
| Landing-page audience | Daily problem and useful result | Detailed guide |
|---|---|---|
| Students · Researchers | Connect papers, readings, research notes, and source evidence so a report separates observations from interpretation. | Research and learning |
| PMs · Designers | Connect customer evidence, requirements, experiments, and decision reasons so the next product iteration starts with context. | Product planning |
| Team Leads · Ops | Keep procedures, support guidance, recurring decisions, and handover checks usable for the next teammate. | Team operations |
| Solopreneurs · SMBs | Give a helper current brand, product, FAQ, and policy references before they answer a customer. | Small-business operating knowledge |
| Writers · Creators | Connect character facts, worldbuilding, plot choices, source research, and channel tone while distinguishing canon from proposals. | A connected story bible |
| Teachers · Professors | Carry curriculum goals, lesson plans, rubrics, and anonymized reflections into the next term. | Reusable course preparation |
These are the six themes on the dtxWiki landing page. The guides follow the v0.1 features and limitations. A landing-page example describes an intended use; it does not establish that every illustrated action is automated in the application.
How the parts work together
In words: keep original sources in Library, explain the relevant meaning in Context, and link related pages from a hub. Use that hub in a real task. When something changes, check and publish the correction so the next task starts from the current decision. This is a reference workflow, not a set of automatic actions.
| You have… | A useful next step |
|---|---|
| A PDF, spreadsheet, slide deck, or permitted source image | Upload it to Library, inspect processing, and compare the relevant extracted text with the original. |
| A reason, conclusion, procedure, or unresolved question | Write it in Context with a clear heading and source reference. |
| Several notes on one project | Create a short hub and use wikilinks and citations so another person can follow the reasoning. |
| A new answer or changed rule | Check the facts, update affected pages, and inspect History after publishing. |
| A teammate or public audience | Check membership or prepare a reviewed public summary using the appropriate sharing action. |
Adopt it in one week
- Choose one recurring task. “Answer a bulk-order question” or “prepare the next lesson” is small enough to try. Follow the closest guide with fictional data first.
- Create the three example pages. Publish the two supporting pages and the hub. Open every link from the reader so the exercise produces a usable reference, rather than only saved text.
- Replace one example with permitted real material. Add an owner, a date, the evidence or source, and a visibly unresolved question. Keep sensitive individual records in their designated systems.
- Use the hub during the next real task. Notice what is missing. Add the answer or decision only after checking it; an unsupported guess is not a useful reference.
- Review at the end of the week. Search for the hub, compare one correction in History, and ask a colleague whether they can find and understand the current published guidance.
Browser Save draft keeps editing unpublished and does not itself submit a Review item. Human Editors and Owners can publish directly; explicitly submitted proposals have a separate review workflow. Workspace publication and Publish to web are also separate. The individual guides explain which readers should receive which material.
What these workflows depend on
Research still needs a person to verify evidence; product and operations still need someone to make and carry out decisions. Business records do not become a POS or automated customer messaging system. Story references do not automatically validate continuity or schedule posts. Course preparation does not become an LMS, a grading engine, or student-record processing.
You can use Search and the linked pages yourself. Ask can retrieve permitted evidence; without an answer provider it returns an evidence summary. An MCP-connected AI tool is optional and requires explicit credentials and workspace access. The example prompts ask for suggestions or comparisons in the conversation, followed by human checking. They do not require write access or a paid AI service.