Knowledge Sources: Give Mailtwine the Knowledge Behind Your Business

A useful email reply often depends on information that never appeared in the email. The answer might be in your product documentation, an engineering decision, a support guide, or a note explaining why a feature behaves a certain way. Finding that context is often more work than writing the reply.

Knowledge Sources brings that context into Mailtwine. Connect your websites and reference material, choose the Streamient projects your inbox should use, and give AI replies a foundation in what your team actually knows. It is a major expansion of what Mailtwine can do: the knowledge behind your business can now help shape the work happening in your inbox.

You manage sources in one central place. Add a page or website, let Mailtwine index its content, and make it available for relevant questions. Connect Streamient as your memory and knowledge source provider to draw on the project knowledge you already maintain there. You can read from all accessible Streamient projects or select the ones that matter to your work.

The real power comes from combining those sources. A public guide can explain the supported workflow. A current implementation note can add an important limitation. A recorded product decision can clarify the reasoning. Mailtwine can search for the issue raised in the email, select relevant passages, and use that evidence when preparing a reply. Your team spends less time gathering context and more time resolving the question.

Streamient makes the potential reach much larger. It brings notes, decisions, saved URLs, and reusable memories into shared, searchable project context. Knowledge captured during development or research can become useful again when a customer asks a question weeks later. Instead of manually copying background into every drafting request, you connect the appropriate projects once and keep the underlying knowledge useful over time.

Imagine you support two products through two email addresses. One inbox handles Helpmonks questions; another handles Razuna. In Streamient, you maintain separate projects containing each product's documentation, decisions, and development notes. In Mailtwine, you assign the corresponding sources to each address.

“Can I share these photos with an external client, protect the link with a password, and have it expire?”

For a question like this arriving at the Razuna address, Mailtwine can search the relevant sharing knowledge rather than pull unrelated information from the Helpmonks project. If your connected material establishes password protection and expiration, those details can support the draft. If it does not establish whether recipients need an account, the reply should leave that point unconfirmed. A note about a planned improvement should not turn into a promise that the feature already exists.

This is how a broad library becomes useful in a specific conversation: the address provides context, the question narrows the search, and the retrieved evidence guides the answer. If someone emails the Helpmonks address with an explicitly Razuna-related question, Mailtwine can recognize the different product and consult the central default sources. You can see when that fallback was used.

Each address can inherit your central defaults or have its own selection of Streamient projects and custom sources. The Knowledge & AI card separates Info, Reading, and Writing: define a custom product name, aliases, and instructions; choose reading sources with searchable tags; and decide whether sent replies should contribute reusable memory to a selected Streamient project. Memory writing is optional, and it is skipped when the question clearly belongs to a different product than the address.

Relevant sources are only part of a good answer. Mailtwine also checks draft claims against the selected evidence, attempts a correction when necessary, and qualifies unsupported details. Previous emails and automatically learned replies are treated as historical context rather than independent proof of current product behavior. These checks help reduce unsupported claims; they do not replace your judgment or the need to maintain accurate source material.

We have also expanded global email search. The former six-result ceiling is replaced by 15 matching email threads per page, with Previous and Next controls, a page count, and the total number of matching results. Results remain grouped by thread and ordered newest first. Compact pagination sits beside the matching-results count and disappears when there is only one page. You can also include Trash and Spam when needed.

That makes it easier to work through the indexed matches for a supplier, invoice, or ongoing conversation instead of stopping at a small initial list. Together, better email search and connected knowledge help with both sides of inbox work: finding the conversation and finding the information needed to answer it.

Start with one address and the sources you trust most. Open Knowledge Sources to connect your material and Streamient, then use Mail settings → Accounts → Edit → Knowledge & AI to tailor an address. As your knowledge grows, Mailtwine can draw on more of it while keeping each reply focused on the product, question, and evidence that matter.