Runko Works

Demos

How the assistants are built

The model is the smaller part. Whether an answer can be trusted is decided around it. Every demo on this site is built the way this page shows.

In plain words

What stands around the model

  1. Where the conversation happens

    Channels
    • WebsiteMobile app
    • Messengers
    • Telegram
    • WhatsApp
    • Instagram
    • Next channel
  2. The harness: one pipeline for every channel

    Orchestrator
    • Saves every message first
    • Queues one reply per conversation
    • Builds the turn: instructions, tools, history
    • Records every step

    The core, left as it is

  3. Custom for each business: configured, or coded

    Deterministic workflow
    • Configured
    • Fields to collect
    • Outcomes and their criteria
    • Consent before contact
    • Handover to a person
    • Coded
    • An app of its own, such as ads analytics
    Next step
  4. A model per assistant, replaceable

    AI
    • Claude
    • Claude in the EU
    • GPT
    • Fallback model
    Next model
  5. What the model can reach

    Tools and knowledge
    • Knowledge
    • Search
    • Whole documents
    • Tables
    • Tools
    • Calendar
    • CRM
    • Price calculator
    • Connected
    • Your own system, over MCP
    Next system
  6. Closed GDPR perimeterEU mode, set per client: website, mobile app, hosting, database, models and embeddings stay inside the EU. A processing agreement, export and deletion for every client.
A modular system. What a business needs is configured, coded, or connected. The core is never rewritten for it.
Where do the answers come from?
From the company's documents and tables. Each demo page shows them next to the chat, word for word, so an answer can be checked against its source.
Can it use our own systems?
Yes, live. What a business needs is configured, coded, or connected: a system that speaks MCP, such as an ERP or a shop, brings its own tools, and the assistant asks it while it answers. See Connected over MCP and the parts & orders demo.
What if it sees only a part?
It says so. An assistant never answers from a part of the data without saying so. A list cut short is called incomplete, and a document that did not arrive whole is not judged.
How do I know it works?
Scripted conversations are run against a copy of the assistant, and every report is read in full. Each turn leaves a record: what it looked up, what it did, what it cost.
What does it cost to run?
One to two cents for a simple reply on the model I use, two to four for a full offer check or a CV scorecard.

Under the hood

Data is prepared before it is searched

Each assistant is a folder in a repository: instructions, settings, documents, tables. One script uploads it, a second checks that the running assistant matches it. Nothing is edited by hand on the platform.

Documents are cut into passages of about 300 tokens, along their structure:

  • A heading stays with the text under it.
  • A long table is cut between rows, and every part repeats the header row.
  • A paragraph is cut between sentences, never inside a word.
  • Every passage carries the title of its document and the headings above it. Both take part in the search.

Against a cut by character count (1,200 characters, overlap 150), on the 80 documents of the seven assistants on this site:

Passages that …Cut by charactersCut by structure
exist in total166213
begin in the middle of a word590
end with a heading whose text went to the next passage410
do not know which document they come from860

Facts that are counted or filtered are also held as tables. A script builds each table from its document, so the two cannot disagree. A cell holds only what the document states: no word about a lift is an empty cell, and empty is not "no".

One PostgreSQL database holds all of it: documents, passages with their embeddings, table rows, conversations and the trace of every turn. There is no separate vector store: with 14 to 76 passages per assistant a comparison with every passage is fast, and the vectors stay under the same access rules and the same deletion as their text.

The calls in the figure come from test runs of the demos.

ModelDecides what the answer needs, picks a tool and writes the call itself.

A question in free words

Search

search_knowledge("Error 204 C axis does not reach target position")

  1. Ranked by meaning, with embeddings
  2. Ranked by words: a rare word weighs more
  3. The two orders are merged
The five best passages, each headed by its document and section
Not the whole document. Not every match.

A document that has to be read whole

Document by code

get_document("RQ-220")get_document("EP-01")get_document("OF-A")

  1. The model sees a catalogue: code, title, size
  2. It takes several documents in one step
The full text, word for word
Too long? A table of contents, never a cut text.

Anything counted or filtered

Exact query

query_data(table: listings_sale, where: rooms >= 3, price_eur <= 500000, about: "needs renovation")

  1. The model names a table, columns and conditions, never SQL
  2. The platform builds the query from a checked schema
  3. It runs under a role that reads only this assistant’s rows
Every row that fits, and how many there are
A list cut by a limit says it is not complete.
Every result says what it coversTable "listings_sale": 4 rows match; all 4 are shown, ordered by closeness to "needs renovation".A wrong call comes back with the valid names, and the model corrects it.
AnswerBuilt from what the tools returned.
No fixed search step. The model asks until the result covers the answer.
  • Search does not assemble a document. The query "check offer A against the requirements" returned no passage of offer A among the best ten. So the offer check and the CV scoring take their documents whole, by code.
  • The model never writes SQL. The pattern is the one from What a semantic view really does.
  • Meaning sits on top of the exact filter. For "three rooms or more, up to €500,000, needs renovation" in the estate agency demo the conditions decide which listings fit: exactly four, the same four every time. The wish in free words only puts the closest one first.

Tests start below the model

The free checks come first and cost nothing to repeat. A run with the model is paid and starts with an estimate.

A change

Instructions, data or settings, in the folder of the assistant.

No model · free

Same as its folder?

The running assistant is compared with its folder in the repository. The tables are rebuilt from the documents.

StopsA setting changed by hand

No model · free

Is the data right?

Seventeen statements about the data, asked the way a tool asks, without the model.

StopsA document that is not whole. A cut list that does not say so.

With the model · cents a turn

Does it behave?

A script plays the visitor through the chat interface, then reads the answers, the events and the trace.

  • Scenarios
  • Grounding
  • Red team
  • Repeats
StopsA booking before consent. An invented number. A rule talked away.

A person · reads

Is the answer right?

The report of the run is read in full: every dialogue, the cost, a written assessment.

StopsAn answer that passes every check and is wrong
Live

A short check after rollout.

What can be checked without the model is checked without it. What no check can see is read by a person.
  • Reference results make scoring testable. The four made-up candidates of the CV demo have to score 95, 73, 68 and 14. The three offers have to come out as 11, 7 and 6 compliant requirements out of twelve, with the exact split between deviation, unclear and not addressed.
  • Checks catch the mechanics, reading finds the defects. The checks are patterns over the answer and the trace. One CV scored 68 in one answer and 73 in another, and none of them failed. The fix went into the data: the vacancy now defines each rating for each requirement.
  • Staging, not live. On the live assistants a test run would fill the working inbox, open cards in the CRM, take calendar slots and use up the daily reply limit.
  • Cost is measured. Seven cases with eleven visitor turns: all green, 26 cents. An offer check that searched seven times cost 12 to 16 cents; with three whole documents in one step, two to four.

See it work: offer check, CV scoring, estate agency, parts & orders on a live ERP, or all demos. Each page shows the data next to the chat.

The platform under these demos is my own. It runs live, and it is not open to external clients yet.

As of October 2026.