«Obzhito» is a lifestyle educational media outlet focused on decorating rental apartments and rooms. The project helps make living spaces comfortable and personal through the strategic use of rearranging, lighting, textiles, storage, and items that can be taken when moving.

In a rental apartment, you want to feel at home, but you must take into account the landlord’s furniture, a limited budget, and the need to coordinate any changes. A beautiful interior shown in a photograph may not always suit a specific room or rental conditions.

«Obzhito» helps you figure out what to change first, which solutions are appropriate for the apartment, how much money and effort they will require, and how to restore the original state. Every article considers the future move—what can be packed and reused.

The core idea of the project is to help people furnish their daily lives in a residence that they do not own.

Original materials and feedback. Interviews, surveys, reader photos, and editorial reviews help select topics and refine recommendations. Search queries in Yandex Wordstat supplement this picture.

For the project, we propose using a combination of Feedly, Make, Airtable, and AI. These services gather available materials, standardize them into a common format, and assist in drafting the text. The editor verifies the sources and the applicability of the advice. The decision to publish remains with the human.

1 Collection Machine and Human — Feedly tracks connected sources; Make accepts new entries. The Editor adds the remaining links and reader feedback. 2 Storage Machine — Airtable saves the source materials, links, collection date, and review status. 3 Processing Machine — The AI extracts the practical task, constraints, and suggested category, noting missing data. 4 Review Human — The Editor verifies specifications and application conditions, clarifies cost, and confirms image rights. 5 Draft Machine — The AI prepares the text based on verified cards. Make returns the draft to the database. 6 Approval Human — The Editor refines the text, checks the illustrations, and approves a specific version of the material. 7 Publication Human or Machine — The material is published manually or via a configured Make scenario after the Editor’s approval. 8 Feedback Machine and Human — The system gathers available responses and metrics. The Editor selects new topics and corrects any discovered errors.

Collection and Archiving. The system ingests material from connected RSS feeds and other available channels, checks for duplicates, and creates entries in Airtable.

Initial Analysis. AI identifies the space type, reader goal, dimensions, materials, constraints, and cost information while maintaining traceability to the source.

Categories breakdown. For submission, please choose one of the following categories: «First Evening, ” „Without a Trace, ” „By Location, ” „My Corner, ” „Is It Worth It, ” or „Bring Along.“

Text preparation. After validating the data, the AI proposes a title, solution options, and a draft post. Unclear information remains flagged.

The editor selects topics that help solve a specific everyday problem. They verify whether the suggested method is suitable for rented accommodation, what surfaces and dimensions are considered, whether landlord approval is required, and what will happen upon dismantling. When working with products, the individual cross-references the instructions, contents, price, and the date of its verification. If necessary, they organize a practical test and describe its conditions. A successful experience in one apartment is not a guarantee for all others. The editor also selects photographs, verifies authorship, answers reader questions, and approves the final text. Automation assists in preparing the material; a person is responsible for the recommendation and its publication.

Feedly Source Monitoring Organize sources by topics: «Rental Housing, ” „Storage, ” „Light and Textiles, ” „Moving.“ Connect available feeds from selected publications and manufacturers. Save the publication name, link, author, and date. Editor manually adds materials from social media and websites that do not have a suitable collection channel.

Data Transfer and Stage Control Configure the transfer of new materials from available sources into Airtable. Before creating a record, check for duplicates. Send only necessary anonymized data to the AI, return the result to the database, and notify the editor. If an error occurs, save the status as «In Review.» Publication submission is permitted only for versions with the status «Approved.»

Airtable Material Storage For each entry, store the title, original link, image, and its author, room, practical task, category, restrictions, cost, and review date. Separately note the dismantling method, post-move usage, and image publication permission. Use the statuses: «New, ” „Under Review, ” „Approved, ” „Draft, ” „Approved for Publication, ” and „Published.“ The editor records the review and approval.

ChatGPT and OpenAI API Text Processing The editor can use ChatGPT manually; in automated scenarios, the model processes the text via the OpenAI API. The AI suggests rubrics, highlights information, and writes drafts based on verified cards. Links and source restrictions are maintained in the text. It is forbidden to invent prices, product specifications, reviews, or test results. If information is lacking, the service formulates questions for the editor.

Tone «Cozy”—calm, friendly, and concrete. We address the reader informally („you“), explain terminology, and suggest attainable solutions. We take into account limited budgets, existing furniture, and personal habits. We do not judge housing as „incorrect“ nor do we create the impression that comfort absolutely requires new purchases.The texts are built around everyday situations: what is holding things back, what solutions are available, what will be needed, and how to assess the outcome. Instead of making promises that it „will work for everyone“ or „will definitely be flawless, ” we specify the conditions for application.

The visual foundation for media involves photographs of living spaces featuring natural light, clear composition, and discernible details. It is important to illustrate the placement of objects, dimensions, and how the solution is utilized. Floor plans of rooms, close-ups of fixtures, and sequences of action are suitable for analysis.

The color palette can be based on milky white, warm gray, beige, and muted olive. This supports the everyday and tranquil nature of the project.

«Before» and «after» photographs must demonstrate genuine changes. Generated images are labeled as concept visualizations. Third-party photographs are published only with the necessary rights or permission; readers' personal materials are published with their consent.

Before publication, the editor verifies sources, terms of use, current pricing, and image rights. The material must clearly outline costs, limitations, the process for reverting to the original state, and whether the items can be reused after moving.

For the first phase of the project, we propose using a Telegram channel for short analyses, curated selections, and reader questions. Automatic distribution is only possible after the text and illustrations have been approved. If a recommendation requires clarification, the material remains under review.

After the publication goes live, we will collect questions and feedback regarding the outcome. If an error is found, the editor will correct the material and indicate what has changed. This feedback helps us plan future topics.