Analysis updated 2026-05-18
Let Claude Code analyze screenshots or diagrams even when its own model can't view images.
Batch-analyze a whole directory of images in sequence.
Generate or edit images by falling back to other API providers automatically.
Configure multiple AI channels with priority order so requests retry on the next provider if one fails.
| hellowind777/hello-multimodal | 13127905/deep-learning-based-air-gesture-text-recognition- | 6xvl/paralives-plugins-index | |
|---|---|---|---|
| Stars | 15 | 15 | 15 |
| Language | Python | Python | Python |
| Setup difficulty | easy | moderate | easy |
| Complexity | 2/5 | 3/5 | 2/5 |
| Audience | developer | developer | general |
Figures from each repo's GitHub metadata at analysis time.
Requires copying a config template, adding API keys, and symlinking into the Claude Code skills directory.
HelloMultimodal is a plug-in skill for Claude Code, the AI coding assistant. Its job is to fill a capability gap: when the AI model currently in use cannot analyze images or generate new images, this skill routes those requests to a different model that can. The user configures which models and API keys to use in a single JSON file, and the skill handles the rest automatically. For image analysis, the skill can examine a single screenshot, diagram, or document, or it can scan through an entire directory of images and analyze each one in sequence. If one configured AI service fails or does not support the requested operation, the skill automatically tries the next one in priority order without requiring the user to intervene. For image generation, the skill connects to image-generation APIs and includes a multi-level fallback system. It tries several different API endpoint formats in sequence, and within each format it attempts a full request first and then a simplified one if the full one fails. This makes it compatible with a wider range of providers and local proxy setups. Additional options include generating multiple images in one call, editing an existing image using a reference photo, choosing a thinking budget level that controls how much reasoning the model applies to complex compositions, and setting a seed value for more reproducible results. The project requires no third-party Python packages beyond the standard library. Setup involves copying a template configuration file, filling in API credentials, and creating a symbolic link in the Claude Code skills directory. The configuration file lets you define multiple AI channels with different models and API keys for vision tasks versus generation tasks, and assign each channel a priority so the fallback order is predictable. The repository is compact: two Python scripts handle all the logic, and a SKILL.md file provides Claude Code with a decision guide for choosing the right parameters when it calls the skill on your behalf.
A Claude Code skill that routes image analysis and generation requests to other AI models when the active model can't handle them.
Mainly Python. The stack also includes Python.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.