VITON13 / Focused service
A knowledge assistant that answers from your documents, cites them and says when it does not know
A RAG knowledge assistant in 10–18 working days, from $113 AI-assisted / $173 human-led: we index up to 200 approved documents from up to 3 sources, apply access rules for up to 3 user groups, test the answers against 50 questions with expected answers and launch it in 1 interface, where every answer cites its source and a clear “not enough information” replaces guesses. The price includes 2 revision rounds on the deliverables. Model and hosting fees are paid to the provider.
Get an estimate in 1 working day↘Written estimate within 1 working dayIn short
What does RAG knowledge assistant include?
RAG knowledge assistant starts at USD 113 with AI-assisted production or USD 173 with a human-led route; typical delivery is 10–18 working days.
Includes 2 revision rounds on the deliverables; VITON13 confirms the scope in writing.
What you get and on what terms
RAG knowledge assistant
AI-assistedfrom $113
Human-ledfrom $173
- Timeline
- 10–18 working days
- Revisions
- 2 rounds on the deliverables
- Included
- only the deliverables listed in the agreed package
- Not included
- external costs and work outside the agreed scope
- From you
- brief, current materials and any access needed for this scope
- You receive
- listed deliverables plus a written completion note
Prices in USD; other currencies are approximate. Scope, revisions and dates are written into your estimate.
Who it is for
A team whose staff or customers keep asking questions already answered somewhere in policies, manuals or a help centre, and who need answers they can check.
What is included
- Up to 200 approved documents indexed from up to 3 sources, such as PDFs, a shared drive or a help centre
- Access rules for up to 3 user groups, so nobody gets an answer from a document they may not read
- 50 test questions with expected answers, run before launch and kept for later checks
- 1 interface, a web chat or a messenger bot, with a source link under every answer
- 1 update schedule that re-reads changed documents and drops removed ones
Not included
- Model, vector database and hosting fees (paid to the providers)
- Scanning, OCR clean-up or rewriting of documents that are hard to read
- Actions in other systems, such as creating tickets (that is an AI agent)
- Any promise of accuracy beyond the tested questions
How the work runs
- 01
We agree the sources, who may see what, and the questions people ask most often.
- 02
We index the documents, split them sensibly and wire in the access rules.
- 03
We run the 50 test questions, fix wrong or unsupported answers and set the “not enough information” threshold.
- 04
You test it with your team; after the revision rounds it goes live with the evaluation report and the update schedule.
Example
Pilot / new service — first client cases in progressHow pilots work
No RAG assistant case is published yet. The first client cases are in progress.
Related guides
All guidesAI automation
How much does a RAG chatbot for company documents cost, and should you build or buy?
A RAG chatbot on company documents costs roughly $4,000–$12,000 for a one-source MVP and $15,000–$80,000 for production scope, according to 2026 price guides, plus about…
Read the guideAI automationHow much does an AI chatbot for a website cost in 2026, and when does it need a human?
Published 2026 price guides put a scripted website bot at about $2,000–$15,000 and an LLM chatbot that answers from your content at $24,000–$60,000 or more, plus…
Read the guideAI automationHow to add AI to an existing website or app without rebuilding it
Usually yes: an AI feature plugs into your current site or app through a server-side route. The common four are an assistant, search over your content, AI inside an…
Read the guideYOUR NEXT PROJECT · DEVELOPMENT FROM $13
From an idea to a clear plan.
Explore the scope and starting cost, then talk directly with VITON13. Your selected service stays with the conversation. Written estimate within 1 working day.
Let’s make it yours.
Explore our work or bring your own idea. It starts with a conversation.
Concept↗
Concept site · WebGL · English · 2026HYPERSOLID
Scope and pricingYou can choose later+
01 / Scope & estimate
RAG knowledge assistant
Starting budgetFrom $113
Turn approved documents into cited answers with access rules, freshness controls and a clear ‘not enough evidence’ response.
Inside this scope
- Up to 200 approved documents indexed from up to 3 sources, such as PDFs, a shared drive or a help centre
- Access rules for up to 3 user groups, so nobody gets an answer from a document they may not read
- 50 test questions with expected answers, run before launch and kept for later checks
- 1 interface, a web chat or a messenger bot, with a source link under every answer
- 1 update schedule that re-reads changed documents and drops removed ones
- Timeline: 10–18 working days
- Revisions: 2 rounds on the deliverables
- Included: only the deliverables listed in the agreed package
- Excluded: external costs and work outside the agreed scope
- Input required: brief, current materials and any access needed for this scope
- Handoff: listed deliverables plus a written completion note
A starting estimate for the selected scope. We confirm deliverables, timing and the final price in writing before work starts. External costs are separate.
What you can count on
- Written estimate within 1 working day, sent to the contact you choose.
- Scope in writing before work starts: deliverables, revision rounds and what is not included.
- We work in four languages and show prices in your currency.
- No account needed. Attachments are encrypted in this browser before upload.
Send a detailed brief or attachments
Development from $13
AI reduces repetitive production time. VITON13 passes that efficiency back to the client while human review remains part of every delivery.
Order with VITON ID
Read the full guide before ordering
Source and permission model — The maintenance reality of RAG knowledge assistant after launch
Plan RAG knowledge assistant from the first working Source and permission model through Retrieval and answer pipeline to an operable Citation and evaluation report. The guide orders dependencies, checks and ownership before production begins.
Open the article↗Questions people ask before buying
01What does citing its source mean in practice?+
Every answer shows the document and the passage it came from, as a link. If the documents do not contain the answer, the assistant says so instead of inventing one.
02Can it use confidential documents?+
Yes, with access rules: each user group sees answers only from the documents it may read. Where the data is processed is agreed in writing before building.
03How do you know the answers are right?+
We do not assume it. The 50 test questions with expected answers are checked before launch; the report shows which pass, which fail and why, and you can run them again after any change.
04What happens when a document changes?+
The update schedule re-reads changed documents and removes deleted ones, so answers follow the current version. The date of the source is shown with the answer.
05Is this the same as a chatbot on our website?+
Not quite. This assistant answers from your internal documents with access rules; a chatbot for site visitors has its own page.
06Which AI model does it use?+
The one agreed in the written scope, chosen for your language, your data rules and your budget. The provider bills its usage to your account.
What clients say
More in this category
AI web application development
Turn a useful AI workflow into a secure product with owned data, review points and measurable output.
↗
AI automation services
Replace one measurable manual bottleneck with a monitored AI-assisted workflow, explicit approvals and a documented fallback.
↗
AI agents for business
Give an agent a narrow job, approved tools, verifiable data and escalation rules instead of an unrestricted chatbot role.
↗
Business website development
Launch a fast, manageable website with clear service routes and enquiry tracking.
↗