AI Image Combiner is an AI image compositing tool designed around a "dual reference images" workflow, whose core purpose is to have two images each take on a clear role and then generate a unified new composition.
I. Purpose and Basic Approach The user uploads two reference images: Image 1 is used to anchor the subject—a clear face, pet, product, or character; Image 2 is used to provide the direction being borrowed—scene, clothing, lighting, color tone, texture, or composition. The tool understands the relationship between the two images through the prompt and generates a new image rather than layering and overlaying the original pixels. As a result, pose, perspective, shadows, clothing, and details may all change. If precise placement, masking, or preservation of original pixels is needed, the material recommends using a manual editor.
II. Main Capabilities According to the official introduction, the tool can be applied to a variety of creative directions, including: pet astronaut portraits, fantasy avatars, warm-toned studio portraits, product lifestyle scenes, original fantasy rangers, anime city stories, and more. The tool displays the currently available models, and users should select a model marked as Available before starting. The models mentioned in the material include: GPT Image 2 (suitable for high-fidelity image input and precise editing), Nano Banana 2 (suitable for rapid creative iteration), and Gemini 3 Pro Image (suitable for complex instructions and professional high-resolution assets). The model descriptions are official recommendations, and actual availability depends on the live workspace.
III. How to Use (Three Steps) STEP 01 Upload two reference images: choose a clear person, pet, or product for Image 1, and add the scene, lighting, clothing, or visual direction you want to borrow for Image 2. STEP 02 Tell the AI the role of each image: write a focused prompt explaining what in the first image must remain recognizable and what can change based on the second reference. STEP 03 Generate, check, and download: after running the tool, check the face, paws, hands, labels, shadows, and small details, and download only after confirming there are no issues. The material also mentions that you can read the full "How to combine two images with AI" guide (the page does not provide the specific linked content).
IV. Applicable Scenarios Suitable for pet owners, gift creators, profile creators, cosplay enthusiasts, storytellers, freelancers, founders, e-commerce sellers, marketers, small brand teams, tabletop gamers, authors, concept creators, social content creators, illustrators, and anime fans, among others. Typical uses include: placing a pet into a spacesuit scene, placing an adult portrait into a fantasy environment, borrowing a warm studio atmosphere for an avatar, placing a product into a lifestyle scene, and generatingOriginal fantasy characters, bringing character portraits into an anime city atmosphere, etc.
V. Usage restrictions and notes
- Generated results may vary even with the same two reference images and prompt;
- Manual review is required for identity, anatomy, hands, claws, product geometry, labels, text, logos, perspective, shadows, and unwanted symbols;
- The tool can assist creative work, but it cannot guarantee complete similarity, authenticity of packaging information, legal permission, or reproducible commercial results;
- Uploaded files and generated files expire after 24 hours; daily credits are used first;
- The official images are OpenAI demos owned by the project; the specific model version was not retained, and they can only serve as directional examples, not as a cross-model benchmark or a promise of reproducibility.
VI. Content not clearly stated in the materials Regarding the maximum number of images that can be merged, supported file formats and sizes, whether an account is required, the privacy policy for uploaded images, and whether they can be used for commercial projects, etc., the official text lists the questions in FAQ form but does not provide specific answers, and none are currently available.
