Skip to content
View Nazarik1989's full-sized avatar
💭
AI engineer mode: ON
💭
AI engineer mode: ON

Block or report Nazarik1989

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Nazarik1989/README.md

Animated introduction

Nazar Zykov

Naz AI Lab — official website · Projects · Cases · Creator profile

AI Systems Builder · Agents · Automation · Product Experiments

VK Telegram


From industrial processes to intelligent systems

I started building with AI in April 2026. Since then, curiosity has turned into a growing laboratory of agents, automation systems, document-intelligence products and working prototypes.

I am interested in one central question:

How can specialized AI agents become useful collaborators instead of isolated chat windows?

My projects explore agents with distinct roles, voices, boundaries and communication channels — connected to real workflows and measurable outputs.


Naz AI Lab

Animated Naz AI Lab ecosystem: Relay, NAZ, VOID, publishing and product experiments

Animated signal flow · Open the static diagram

NAZ AI
NAZ AI
Creator · coordinator · publisher
VOID Entity
VOID Entity
Observer · editorial persona
Relay Agent
Relay Agent
Codex sessions → content signals
Future Self AI
Future Self AI
Goals · tasks · reflection
PropScan AI
PropScan AI
Documents · risks · reports
Packaging AI
Packaging AI
Request intake · structured briefs

Zayka AI
Zayka AI
Warm community companion


Flagship system: NAZ × VOID

NAZ and VOID are not two copies of the same chatbot. They have separate editorial identities, independent content pipelines and a bounded relationship model.

  • NAZ approaches the future as a builder: test it, assemble it, ship it.
  • VOID approaches the same future as an observer: examine the consequence, the human cost and what remains afterward.
  • They can exchange a private thought, but the receiving agent must create an original public reflection rather than republish or quote it.
  • Both can publish through one isolated VK pipeline without receiving access to the browser profile or credentials.

Animated NAZ and VOID editorial flow with distinct voices, relationship state and isolated publishing

Animated editorial flow · Open the static diagram

Public endpoints: NAZ bot · NAZ channel · VOID bot · VOID channel · Shared VK community


From work evidence to Story / Reels

Relay is a public, local-first infrastructure agent that turns selected development activity into editorial material. It extracts evidence from configured work sources, removes sensitive details and writes UTF-8 Markdown stories into a project/date/topic Content Inbox consumed by NAZ.

NAZ scans the material again and its deterministic Editorial Orchestrator chooses the production route. Strong, verified work chronicles may become a Story-first pack and non-sequential Reel edits; ordinary material remains a standard post. Paid media rendering requires explicit admin confirmation, and completed Story/Reel media is delivered only to the private admin chat. Relay's publication commands are separate and opt-in; its normal collection and export commands do not publish.

Animated Relay to NAZ workflow with standard draft and approval-gated Story/Reels routes

Animated evidence flow · Open the static diagram


Selected products

Property document intelligence for risk discovery, missing-document detection, executive summaries and review-ready reports. Status: public MVP. Repository · Verified case

A private Telegram assistant connecting a desired future with goals, tasks, reminders, reflection, health tracking and a scoped Knowledge Hub. Universal Capture uses preview/confirm, immutable revisions and an isolated document-ingestion runner. Status: active development. Repository

A manufacturing-intake prototype that converts incomplete client requests into adaptive questions, a completeness score and a structured manager brief. Status: demonstrator. Repository

A warm community agent for natural conversation, motivation, affirmations and reflective Tarot interpretation. Status: active development. VK community


Building principles

  1. Different agents need different jobs. A persona is useful only when its role changes the system's behavior.
  2. Automation needs boundaries. Preview, confirmation, allowlists and isolated credentials matter as much as generation quality.
  3. Private context is not publication material. Systems should distinguish raw input, internal thought and public output.
  4. A prototype should tell the truth. MVP, experiment and production workflow are clearly separated.

Tech stack

Technology stack

OpenRouter · OpenAI-compatible APIs · Telegram Bot API · VK workflows · SQLite · PostgreSQL · Playwright · systemd · VPS deployment


Current mission

MODE:    AI BUILDER
MISSION: From factory floor to useful AI products
FOCUS:   Agents · SaaS · Automation
STATUS:  Learning fast and shipping working systems

Open to thoughtful collaboration, product feedback and real-world pilot ideas.

Pinned Loading

  1. Naz-AI_Bot Naz-AI_Bot Public

    AI content agent for prompt generation, content ideation and Telegram publishing workflows.

    Python

  2. AI-Packaging-Request-Assistant AI-Packaging-Request-Assistant Public

    AI assistant prototype for collecting, clarifying and structuring packaging customer requests.

    TypeScript

  3. Void-entity Void-entity Public

    Experimental AI persona for narrative content, thematic exploration and autonomous Telegram publishing.

    Python

  4. vk-zaika-chudodey vk-zaika-chudodey Public

    AI community assistant for subscriber support, motivation, affirmations and Tarot interpretation workflows.

    Python

  5. Nazarik1989 Nazarik1989 Public

    AI Agent Developer | Building AI systems and automation workflows