Getting Started with Inteleto: Your First Grounded Answer in Under 90 Seconds
In short
Upload documents, open the blueprint workspace, and ask a question. Inteleto retrieves the matching passages, answers from them, and cites each one. You save the blueprint afterwards — configuration comes after the first run, not before it.
Run first, configure later
Most agent tooling asks you to declare everything up front: personas, tools, instructions, output schemas. You spend an hour configuring before you learn whether the thing can answer a single question about your documents.
Inteleto inverts that. Create a blueprint, drop in documents, and ask. The blueprint is what you save once you have an answer worth keeping, not a form you fill in beforehand.
Step 1 — Add your documents
A knowledge base is a scope you can retrieve over. Upload PDFs, scanned documents, or exports from wherever your material lives. Ingestion chunks each document, embeds the chunks, and indexes them for both vector and keyword retrieval.
Scanned and image-only PDFs go through OCR automatically, so archives that were never text-searchable become retrievable without a separate preprocessing step.
Step 2 — Ask a real question
Open the blueprint workspace and ask something you already know the answer to. This is the fastest way to judge retrieval quality: you can check the citations against your own knowledge immediately.
Each answer arrives with the evidence attached — the specific chunks that supported it. If the retrieved evidence does not support a claim, the answer does not make the claim.
Step 3 — Promote what worked into a task
When a question turns out to be one you will ask repeatedly, promote it to a task. The task keeps the prompt, the scope it retrieved over, and the output contract it should satisfy.
Tasks and chat run through the same execution path and produce the same Run record, so anything you can verify in chat behaves identically once it is scheduled or called through the API.