AI glembee

Context We have an internal template library and a section type library. We need to generate images according to the content — for a specific section, niche, and size. We will provide an example of a section showing what specific images are needed, and you will generate them according to the context, style, and size. You will also have access to our internal tools. Scope of work: in total, there should be about 150 high-quality images — consistent in style and palette. Who we are looking for A person with a professional background in design/photography/color correction who generates images with an understanding of the principles of quality visual content. What needs to be done Generate images for the template and section library — according to the content, niche, and size of the specific section. Maintain a consistent style — all images in one style and palette. Cover various niches and industries — from cosmetology and spa to production shots of factories. Select the right tool for the task — understand where the best results are achieved for a specific type of image. What is important We care about not the theoretical difference between types of photography (which everyone knows), but the ability to see and fix problems in the generated image: harsh shadows, overexposure, incorrect focus, "dead" color. You should feel when the image is "off" — and know how to fix it through prompts. Working with focus and depth of field — a conscious choice of the focal point, not a random result of generation. Attention to tone and color — understanding how to keep the image in the desired tonal range at the generation stage or fine-tune the necessary shades in Lightroom, Photoshop, or another editor. Confident use of AI image generation tools — you already work with them daily and understand the strengths of each: where which software gives the best result for a specific type of image. How we work Payment is hourly, discussed individually. Payment for necessary AI subscriptions. How to respond Send a portfolio with examples of your work. Tell us about your most challenging client case — or a case similar to our task. What was difficult and how did you solve it? Add a few sentences about yourself. Important: we read all applications, but prefer those with a brief explanation of why you are particularly interested. Thank you!

Developer for creating AI bots - multi-agent system and automations in AI application. What is needed: proficiency in n8n; experience integrating OpenAI and other AI models via API; working with webhooks, REST API, and Supabase; ability to create reliable AI workflows and automations. Create multi-agent systems from scratch or with a framework AI Agents, MCP, RAG

We have gathered the design and structure for our product ourselves, but it looks bland to us, and we would like to make it more vibrant, refined, in short, more expensive-looking) The order of the blocks remains the same as it is now, but we would like to work on the design Add visual elements, there are some photos of the product author that we could incorporate We had a design done before, but only the video remains (I will attach it as well) It can be used as a reference In general, we want it to look not like a neural network; the main task is to achieve this I am attaching the website https://www.prima-numerology.com/ (the design we put together ourselves), but it turned out to be very strict and template-like In the redesign, do not focus on the product author; her photo can be diluted a bit, but it would be nice to add neutral visuals It would be great if you have top design skills and know how to achieve this vibrancy Google Drive with files https://drive.google.com/drive/folders/1yNZDjj6FaIzr1ESd_nwRB7_v5YEnPXhI?usp=sharing There are photos and an example of one of the designs that were made for us This is the color palette; our brand style is built around it (without changing the colors)

Order: creation of an AI application/bot on Telegram for generating photos and videos I am looking for a programmer or a team to create an application functioning as a bot on Telegram. The main task of the bot will be to generate new photos and short videos based on photographs sent by the user. Main operating principle The user activates the bot on Telegram, sends their photo, and then selects one of the available transformation options. Sample process: The user activates the bot. Accepts the terms and conditions for photo processing. Uploads one or several photos. Selects a style, template, or type of material they are interested in. Makes a payment or uses previously purchased credits. The bot sends the photo to the appropriate AI model. The user receives the generated photo or video directly on Telegram. Possible photo generation features The bot should allow for the creation of, among others: (examples to be determined) professional business photos, photos for CVs and LinkedIn, studio portraits, photos in various outfits and locations, sporty, business, elegant, and humorous stylizations, profile pictures for social media, photos in a chosen style, realistic changes of background, hairstyle, facial hair, clothing, or surroundings, ready packages of several photos generated from one photograph. Possible video generation features In later stages or from the beginning, the bot should allow for: (examples to be determined) animating the uploaded photo, creating short videos with a moving character, generating promotional videos, creating a speaking character based on a photo and text, adding voice or sound background, creating short materials for TikTok, Instagram, and Facebook, generating several variants of the video to choose from. Selection panel The user should be able to choose options using buttons in Telegram, including: photo or video, selected style, number of generated materials, vertical, horizontal, or square format, standard or premium quality, optionally additional text instructions. In the future, the administrator should have the ability to easily add new styles and templates without rebuilding the entire system. Payments Payment should be made directly in the application or through a convenient payment link. Considered model: payment for a single generation, purchase of credit packages, monthly subscription, free credits to start, discount codes, different prices for photos and videos, higher price for faster generation or better quality. The system should automatically calculate credits and block generation in case of insufficient funds. Preferred payment methods: payment card, BLIK, fast transfer, Apple Pay and Google Pay, Telegram Stars, if it complies with Telegram's rules and sales model. The programmer should propose the best payment solution, considering sales in Poland and the possibility of later entering other markets. Administrative panel A simple admin panel will be needed, where one can: check the number of users, check the number of generated photos and videos, manage prices and packages, grant or revoke credits, add promotional codes, block users, check transaction history, control costs of using AI models, manage styles and templates, set generation limits, analyze basic sales statistics. Integration with AI I do not assume creating my own AI model from scratch. Initially, the system can use available APIs for generating and editing photos and videos. The programmer should: propose suitable AI models, compare their costs, evaluate the quality of generated materials, prepare the possibility of changing the AI provider in the future, ensure task queuing and error handling, optimize the costs of each generated material. It is important that the system is not permanently dependent on one API provider. Security and user data The application will process users' photos, so it is necessary to secure the data appropriately. The system should consider: the user's consent to process the photo, terms and conditions and privacy policy, compliance with GDPR, automatic deletion of photos after a specified time, the possibility of deleting data at the user's request, securing files and payment data, restrictions on generating prohibited content, confirmation that the user has the right to use the submitted photo. The bot should not allow the creation of materials that violate the law, impersonate others, create illegal content, or use images without consent. Technical requirements I expect: creation of a Telegram bot, creation of a backend handling users, credits, and generation, integration with AI model APIs, integration with the payment system, creation of a database, preparation of an admin panel, deployment of the application on a server or in the cloud, system security, preparation of documentation, transfer of source code and rights to the completed solution, ensuring the possibility of further development of the application. The system should be prepared for a larger number of users and the ability to handle several generations simultaneously. First MVP version The first version may include: user registration via Telegram, uploading one photo, choosing from several ready styles, generating photos through one selected API, credit purchase, online payment, sending the result via the bot, basic admin panel, history of generated materials and payments. After testing the MVP, further development is planned for videos, subscriptions, referral programs, additional styles, and language versions. Additional features for later implementation referral and commission system, partner panel, English and German versions, separate web application, automatic content publishing on social media, the ability to generate photos for companies and employees, business packages, white label, API for external clients, notification system for promotions, saving user's favorite styles. What I expect from the contractor Please send: technology proposal, estimated cost of creating the MVP, expected cost of maintaining the system, cost of using AI for one photo and one video, examples of previous implementations, information on whether the contractor will also handle the deployment, information on later support and development, proposals for implementation stages, information on which elements may generate the highest costs or risks. Settlement can be staged, after the completion and acceptance of subsequent parts of the project. I am interested in creating a simple, intuitive, and scalable product that can be quickly launched in a test version and then developed for the Polish market.

1. Title Motion Designer: Lottie Animation of Face Analysis for Landing (Web, Vector)2. What We Do One vector animation is needed for the hero section of the landing page of a web product (AI face analysis, playful format). Plot: a stylized wireframe face (linear, semi-transparent, drawn by an artist — not a photo). Landmark points (eyes, nose, jawline, eyebrows) sequentially "sit" on it, with a light grid built between them. In the end, 2–3 small score cards gently "fly out" from the face (e.g. Curiosity 88 / 100). Behavior: build-in on load (the grid builds up, cards appear, ~3–4 s) → then a quiet seamless loop (~8–12 s): the grid barely "breathes," cards gently pulse. No sharp movements.3. Style Aesthetics: editorial, clinical, grey — calm light "studio air," thin lines, soft light from within. Tone reference — QOVES face-analysis (premium "medical-analytical report" × magazine spread). Important — tone: this is a playful vibe analysis, NOT a medical scan and NOT facial recognition. No IDENTIFIED frames, crosshairs, red, "security" HUD, cyber-neon. Points and annotations should feel interesting and light, not like biometric surveillance. Palette and typography — strictly according to the extract at the end of this document (adherence to the palette is a condition for acceptance). References I am adding: QOVES aesthetics + 2–3 examples of Lottie in grey-clinical / wireframe style + a screenshot of the hero where the animation will be (there are glass cards nearby — the animation should coexist with them, not compete).4. Technical Requirements Format: Lottie JSON, transparent background. File size: ≤ 150 KB (ideal ~80–100 KB). 30 fps. No raster images inside — everything is vector (no embedded PNG/JPG). Minimum masks and large gradients; no effects that Lottie does not render natively (complex blur, some merge-paths). Must test export specifically in the web player (lottie-web / LottieFiles preview), not just in After Effects. Mobile: plays correctly and reads on real iOS Safari and Android Chrome; on a narrow column, the face/grid is not cut off, cards do not exceed boundaries; smoothly on budget Android. Artboard that scales cleanly in the range of ~360–700px width. Delivered with a reduced-motion frame — static final state (built grid + cards) for prefers-reduced-motion.5. Deliverables + Rights The contractor delivers: Lottie JSON for the landing. Reduced-motion static frame (SVG or PNG). Editable source file — After Effects project file (.aep) with all layers and labeled groups. (Optional / bonus) "quiet" light version for the login page from the same layers — if you stay within budget; otherwise, it will be moved to the next stage. Rights: full transfer of exclusive property rights + unlimited commercial use. The contractor confirms that all elements are original or licensed for commercial use. Fonts in the cards — free Google Fonts (Hanken Grotesk, JetBrains Mono), no paid font traps.6. Process Style frame on the first day — one static frame of the future animation for style approval. I confirm the same day. After my "okay" — animation. 2 rounds of revisions included in the price. Payment in stages through Safe agreement: part after style frame approval, the rest — after acceptance of final files. Important quick start: style frame on the first day, so respond only if you are truly available now.7. Budget and Timeline Budget: $150 for the full scope with deliverables and rights (fixed). Timeline: 4 days from the start.8. What to Send in Response 1–2 works specifically in Lottie / motion (not just static illustrations). Short comment: how you see this style according to our reference. Confirmation that you can start immediately and provide a style frame on the first day. Responses without Lottie examples and without a style comment will not be considered.Palette Extract (to be given to the selected contractor) This is all the color/typography information needed by the artist. Full CSS/repository is not provided. Background: neutral light #f2f5f7. Grid lines, face outline, text: grey "ink" tones —#1f2730 (main), #56606c (secondary), #8a93a0 (lighter). Single accent — only on landmark points: grey-slate #505c65 (deeper variant #3b4651). No other accents should be used. Prohibited: any other colors, saturated gradients, rainbows, neon, red, heavy shadows. Shadows — only soft, low opacity, with a cool tint. Score cards (trait-reveal): semi-transparent "glass" tiles (frosted, light blur, thin light edge); large number — thin weight 300, elegant, not bold; label — small mono-font, ALL CAPS, spaced tracking. Fonts: Hanken Grotesk — trait names, labels (weight 300–500). JetBrains Mono — small labels, indices, digital signatures (e.g. / 100). Tone in one line: calm almost white studio air, illuminated from within; thin lines; one grey-slate accent touches only the grid nodes.