Internal automation
We remove the manual bridges between spreadsheets, inboxes and systems.
We take the process apart step by step and leave one route for the data: where it comes from, how it is checked, what happens to it and where the report lands. We connect the services you already use, so the team barely has to relearn anything.
A fit when the team spends half a day moving data by hand and still nobody sees the whole picture. Also a fit when a report is assembled by hand every week and is out of date by the time it is ready.
Where to start
You have a specification
Send the document by email. We answer with an estimate in hours, a deadline and the questions it leaves open.
Send it for an estimatehello@tok.moscowYou have no specification
We start by writing one. Questionnaire, a day of briefing, the document in two days, then the work is priced from it.
The Technical specification service- Working hours
- 50–110 hours
- Duration
- 18–35 days
- Hourly rate
- $480
- Price
- $24,000 – $53,000
Stages of work
Drag the edge of any stage. It changes width, the neighbours shift, and the deadline and price recalculate.
- Working hours
- Duration
- Price
- 01Measure where the time goes
- 02Data and rules model
- 03Rollout and integrations
- 04Reporting and team training
What we automate
Collecting
Requests, letters, exports and spreadsheets come together in one place with a shared structure.
Processing
Rules, checks and text analysis by a model. Anything doubtful goes to a person, the rest runs on its own.
Delivery
Reports, notifications and records inside the system the team already works in.
Perimeter and services
Data stays in local data centres, payment in roubles or tenge, paperwork to local requirements.
- Yandex Cloud
- servers, queues, object storage
- VK Cloud
- a second site inside the same perimeter
- Cloud.ru
- compute and databases
- YandexGPT
- text analysis without leaving the perimeter
- 1C
- exchange of documents and catalogues
- Telegram
- notifications and simple flows
For teams outside the region: familiar services and payment by any bank card.
- AWS
- compute, queues, storage
- Cloudflare
- edge processing and protection
- OpenAI
- text analysis and drafts
- Stripe
- taking payments
- Slack
- notifications and approvals
- Notion
- knowledge base and tables
The client's own infrastructure: we install on their hardware and nothing leaves it.
- Docker
- build and run
- PostgreSQL
- data
- Local models
- text analysis on your own hardware
- Keycloak
- sign-in and permissions
- Prometheus
- watching how it runs
- S3-compatible storage
- files and backups
What stays with you
Sources
code and settings in your repository
Infrastructure
described as code, rebuilt from scratch
Documentation
how it works and how to fix it
Access
handed to your team