How to Deploy AI for Document Search: Runity's Experience
Runity has described its corporate RAG assistant, which combines search across Confluence and GitLab, checks access to each document, and does not send data to external services. Under the hood are local Qwen models, Qdrant, Temporal and four separate agents, ranging from a general chatbot to a coding helper. The team estimates that infrastructure for such a solution costs 160–200 thousand rubles per month.
AI-processed from Habr AI; edited by Hamidun News
Runity has shared its experience deploying an AI assistant for searching documentation in Confluence and GitLab. The assistant searches the codebase and documents at the same time, respects access rights, and runs in a closed perimeter without sending corporate data to external services.
How does a RAG assistant for documentation work?
The system searches Confluence and GitLab simultaneously, respects access rights, and runs in a closed perimeter without sending corporate data to external services.
How do you deploy AI for document search without the cloud?
According to Runity's experience, the system runs in a closed perimeter, and corporate data is not sent to external services. The AI assistant searches Confluence and GitLab at the same time and takes employees' access rights into account.
What can replace manual documentation search?
Runity built an AI assistant for searching documentation and code. In the company's experience, it solves the problem of finding current documentation and code snippets in hours instead of hours of manual searching.
Key Questions Answered
How does a RAG assistant for documentation work?
The system searches Confluence and GitLab simultaneously, respects access rights, and runs in a closed perimeter without sending corporate data to external services.
How do you deploy AI for document search without the cloud?
According to Runity's experience, the system runs in a closed perimeter, and corporate data is not sent to external services. The AI assistant searches Confluence and GitLab at the same time and takes employees' access rights into account.
What can replace manual documentation search?
Runity built an AI assistant for searching documentation and code. In the company's experience, it solves the problem of finding current documentation and code snippets in hours instead of hours of manual searching.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.