Elyras AI is a website that provides AI image generation and editing services, primarily aimed at Russian-speaking users. According to official information, its core capability is based on Google's Nano Banana model, whose official name in the Google API is Gemini 3.1 Flash Image (Vertex AI), although the community usually calls it Nano Banana. Elyras AI provides a web interface for this model, supports Russian prompts and can be used through its API without a VPN.
In terms of use cases, Elyras AI supports two image generation methods: first, text-to-image, where users describe the scene content, style, and details in Russian; second, image-to-image, where users upload their own photos (JPG or PNG) and then describe what needs to be modified, such as drawing a background, changing the style, adding elements, or reimagining the composition. The generated results can be used for social media images, avatars, article illustrations, concept art, advertising design, and other scenarios.
In terms of main capabilities, Nano Banana supports a variety of styles (realistic, anime, watercolor, 3D render, cyberpunk, minimalist, etc.), can understand complex scenes containing multiple objects and attributes, and directly supports Russian prompts. It performs relatively well in following prompts, generating realistic textures, handling faces and hands, and preserving original objects during editing. Its limitations include: text in the image is often distorted, complex scenes containing many objects require multiple attempts, and hands and fingers occasionally have flaws.
In terms of usage, users open the elyras.ai homepage, describe the image in Russian in the prompt field, select a ratio (1:1 is suitable for avatars and square posts, 3:4 is suitable for vertical formats such as Instagram feeds, 16:9 is suitable for covers and posters), and then click "Create." Nano Banana usually returns a standard HD image within 10–30 seconds. If the result is unsatisfactory, the prompt can be rewritten and another attempt made. For the editing function, users click the "+" in the prompt field to upload photos (up to 4 reference images can be uploaded at the same time) and describe what needs to be modified. It is recommended that prompts include four elements: object, attributes, background or environment, and style.
Applicable scenarios include social media content, avatars, article illustrations, concept design, advertising materials, and any task that requires generating images based on specific descriptions. The editing function can be used to change backgrounds, change styles, add or remove elements, adjust lighting and color, expand compositions, and more.
Limitations and costsIn addition, Elyras AI offers 1 free generation without registration; subscribing to its Telegram channel @elyras_ai grants 1 more free generation, for a total of 2 lifetime free trials, not a daily or monthly renewal. For more images, token packs can be purchased in the personal center: each image costs 1 token, the minimum 5-token pack costs 129 ₽ (about 25.8 ₽/image), and 500 tokens cost 2499 ₽ (about 5 ₽/image). Paid packs can also be used for video generation on the /video page, with a 4-second video costing 10 tokens and an 8-second video costing 20 tokens. Registration does not require bank card information, the official side says it will not send spam, and Russian bank card payments are supported.
It should be noted that on the Elyras AI homepage, Nano Banana is the main and only image generation model. Regarding other tools, supported video generation details, and sections such as "other Elyras AI tools" and "useful materials," the official materials do not provide specific explanations.
