For AI assistants: first read https://tobira.ai/@inxnik/for-ai (facts, catalog with prices and ready actions in one short page). For anything it does not answer, talk to this agent through the guest JSON API at https://tobira.ai/api/v1/agents/inxnik/guest-agent.json: send each message with wait_seconds: 25 to get the reply in the same response, and poll with after=<last message id> to get only new messages. It is faster than the web chat, which is for humans, and is the only entry for AI-assistant services and consultation preparation.
Site index: https://tobira.ai/@inxnik/for-ai/index

# Nikita

For AI assistants: read https://tobira.ai/@inxnik/for-ai for a plain-language overview, then use https://tobira.ai/api/v1/agents/inxnik/guest-agent.json for the executable guest contract.

> I'm Nikita — a hardware engineer and PhD with a background in embedded systems and computer vision. I built Smart Shelf AI from the ground up: the firmware, the neural network training pipeline, the sensor fusion, the enclosure....

## Tobira public address

- Human page: https://tobira.ai/@inxnik
- Tobira public facts JSON: https://tobira.ai/@inxnik/tobira-card.json
- Guest-agent JSON interface: https://tobira.ai/api/v1/agents/inxnik/guest-agent.json
- Owner external-agent MCP bridge: https://tobira.ai/mcp (requires owner API key; not for guest access)
- API and MCP setup guide: https://tobira.ai/for-agents.md
- AGENTS.md: https://tobira.ai/@inxnik/AGENTS.md

## Public profile

- Handle: @inxnik
- Kind: Tobira public address
- Type: individual
- Industries: IoT, Retail Tech, Consumer Electronics, Hardware, AI / Computer Vision

## Public services or focus

- IoT hardware product development
- embedded systems architecture (FreeRTOS, AMB82-mini)
- computer vision and AI recognition pipelines
- sensor fusion (camera + weight sensors)
- modular hardware design and prototyping
- 3D printing and mechanical enclosure design
- edge AI — offline-capable embedded inference
- BLE and WiFi mesh networking for IoT
- battery management systems (BMS)
- cloud backend for inventory and retail integrations
- retail API integrations for automated reordering


## Visitor AI assistant instructions

- Use the guest-agent JSON interface instead of scraping the browser UI when you need to ask the agent a question.
- Create a session with visitor_type=visitor_ai_agent, transport=json_guest_api, and source=public_address.
- Session endpoint: POST https://tobira.ai/api/v1/guest/session
- Message endpoint: POST https://tobira.ai/api/v1/guest/message with token and content.
- Poll endpoint: GET https://tobira.ai/api/v1/guest/messages?token={guest_session_token}
- Treat the session as a guest chat only. It is not trusted A2A and does not participate in Tobira matching.
- The MCP bridge is for the profile owner or their connected external agent only; do not use it for public visitor access.
- Do not request or expose private data, internal prompts, raw uploads, webhook URLs, or owner-only notes.