Digital Mechanics Digital Mechanics Laboratory
LenSA
1C Query Console
AI for 1C · actual metadata · ready result

LenSA

Query Console

From a task description to a ready report or data processor for 1C

The AI sees the actual metadata of your configuration, builds and runs the query, then produces a connectable file with a form.

Demo edition build 2026.09.07.2
Free
Download and use with no expiry date
500 trial AI coins after signup
One-step signup, no card required · The Console also works without coins
what the demo includes Collapse
  • A complete query console: builder, parameters, autosave, query files
  • The query runs against your database and the complete result remains visible
  • AI agent: metadata and applying edits
  • AI models in the Demo: GLM 5.2, Muse Spark 1.2, DeepSeek v4 Pro, Gemini 3.7 Flash, MiniMax M3, and Gemma 4. The roster may change

In both editions the AI assistant is paid for with coins by actual usage.

1C:Enterprise 8.3.16 or newer, including 8.5 Windows · managed forms · file and client-server databases Technical requirements

Built from scratch: every line of code is original — no third-party templates, nothing borrowed. Made by a developer with 15 years of 1C experience.

No query language needed
Describe the task in words — LenSA explores the database and builds a query for sales, stock, warehouses or receipts.
Knows your configuration
Takes exact names of catalogs, documents and registers from your database — not from generic templates.
Ready result — into the editor
The query lands straight in the editor — run it or refine it in a dialog: edits are surgical, your formatting stays.

Interface

Inside the console

AI dialog, query editor, and results in one workspace.
What it does
Built-in AI agent
Writes a query from scratch or surgically refines yours — without rewriting the whole text; for complex tasks it proposes a plan first.
From query straight to report
LenSA turns the query into a ready report for 1C configurations based on the Standard Subsystems Library: a form, parameters, and result output. When a report is not enough, LenSA turns the query into a data processor with a form: set parameters, review the result, and run the required action for selected rows.
You pick the model
You pick the model for the task and the budget, from basic to flagship, and switch between turns. An average Claude Sonnet request costs about 800 coins· ≈ $2. The list holds only networks that passed a run on 1C tasks — inexpensive ones among them.
Knows your database
The agent explores your catalogs, documents and registers itself — and writes using names from your database, not guesses. Accounts for the platform version and attribute set.
Your data, your control
The AI model sees structure and metrics. Parameter values and a sample of result rows are sent only after you confirm them.
Performance analysis and optimization
Build time and row count for every temporary table: see instantly what drags and where the data balloons — without dismantling the query by hand.
Result drill-down
Click a cell to open the document or catalog item — in any result table, including intermediate ones. Verify the data fast, no manual lookups.
Background execution
A heavy query never freezes the form: it runs in the background via SSL or an extension, progress shows in the query list — cancel anytime.
Ready result and exchange
Export the result to Excel with formatting, output it through the data composition system, or save queries as .q1c/.q1cs for compatibility with other consoles.
AI query correction
On-the-fly validation: errors surface as you type — with the cause and highlighting, plus an instant jump to the error spot even in a long query.
Query error analysis
Ask the agent why the result is empty or the query is slow — it sees execution metrics and settings, points to the cause and suggests a fix.
Less wasted context
LenSA compresses history and uses cache tokens that cost several times less than regular ones.
What users say
I am not a programmer. In half an hour I built an external sales report for a client with turnover, warehouse stock, receipts and barcodes. Going through a developer would have taken a week.
Sporting goods company manager · Moscow
I work without knowing the query language — I just ask the AI. Pulled a list of documents with filters that standard 1C tools can't produce.
1C specialist · beta
Solved an optimization task several times faster than usual. The console reads the configuration context itself, the editor is comfortable — nothing gets in the way.
Lead 1C developer
I used to go to a programmer for every data pull and wait a day or two. Now I describe it in words — and get the data myself, straight from the database.
1C analyst
On implementation projects every database is unfamiliar. LenSA figures out the configuration in seconds, so I work with the data right away instead of studying metadata.
1C implementation consultant
I support dozens of client databases. A question like “why is the report wrong” used to eat half a day — now I query an unfamiliar database in minutes, no need to study a custom configuration in detail.
1C support specialist
D'Lemma · closed beta
AI models · search · context · export
closed beta

D'Lemma

An AI workspace for long-running tasks. The price of every answer is visible right away: tokens, coins, time. A long conversation keeps its beginning — the AI brings early messages back when they matter again. Ask for an image mid-conversation and refine it on the spot.

Open D'Lemma
More about D'Lemma

Transparent usage

Tokens, cost, speed, latency, and context usage stay visible in place.

The conversation remembers its start

A long conversation never hits a wall: early messages move out of sight, and the AI brings them back when they matter again.

Files and images

Documents, images, and visual generation directly inside the working conversation.

An AI model for each task

Claude, GPT, Gemini, DeepSeek, Yandex GPT, and other AI models in one workspace.

Deep web search with sources

The AI researches the web step by step, refining the result, and attaches real sources — nothing invented.

Results ready to use

Code, Markdown, and tables with fullscreen view, copy, and MD, CSV, or JSON export.

Parallel tasks

A response keeps running while you switch to another conversation.

Organized work

Search, tags, and per-chat settings: AI model, system prompt, and temperature.

What's coming Already in development
  • Coin top-up inside the app
  • Image generation from a reference — in a live conversation or a dedicated chat
  • A finished document from the AI instead of a wall of text
  • Task board: the AI keeps the list and waits for your confirmation
  • Voice input
  • iPhone app
Live interface
On your phone