Search Results for "image text input"

The ComfyUI-HunyuanVideoWrapper project is a ComfyUI extension that integrates Hunyuan-based multimodal video generation models into node-based workflows. It allows users to generate or manipulate video content by combining text prompts with one or more input images, enabling flexible conditioning of outputs. The system introduces specialized nodes such as text-image encoders that allow multiple image inputs to be referenced directly within prompts. This makes it possible to guide generation using both visual and textual context simultaneously. The wrapper is designed to fit seamlessly into ComfyUI pipelines, enabling chaining with other nodes for advanced workflows. ...

The Multi-Agent Framework: Given one line Requirement, return PRD, Design, Tasks, Repo. Assign different roles to GPTs to form a collaborative software entity for complex tasks. MetaGPT takes a one-line requirement as input and outputs user stories / competitive analysis/requirements/data structures / APIs / documents, etc. Internally, MetaGPT includes product managers/architects/project managers/engineers. It provides the entire process of a software company along with carefully orchestrated SOPs.

...It acts as a standalone wrapper layer that allows developers and creators to integrate experimental features and models without modifying the core ComfyUI codebase. This design makes it easier to rapidly test new capabilities such as text-to-video and image-to-video generation while avoiding compatibility issues with the main framework. The project supports complex node-based pipelines where users can control sampling, conditioning, and frame continuity across generated sequences. It also enables extended video generation by linking outputs between iterations, allowing for longer and more coherent animations. ...

Phone Harness is a lightweight system that lets AI coding agents interact directly with a real iPhone through macOS iPhone Mirroring. It requires no jailbreak, Xcode project, WebDriverAgent, or custom software installed on the phone. Screenshots of the mirrored window provide visual input, while Apple's Vision framework extracts visible text and tap-ready coordinates. macOS CGEvents supply taps, long presses, drags, scrolling, keyboard input, and iPhone Mirroring shortcuts. After performing an action, the agent captures the screen again to verify the actual resulting state. The harness can be installed as an agent skill for systems such as Claude Code or Codex. ...

CogAgent is a 9B-parameter bilingual vision-language GUI agent model based on GLM-4V-9B, trained with staged data curation, optimization, and strategy upgrades to improve perception, action prediction, and generalization across tasks. It focuses on operating real user interfaces from screenshots plus text, and follows a strict input–output format that returns structured actions, grounded operations, and optional sensitivity annotations. The model is designed for agent-style execution rather than freeform chat, maintaining a continuous execution history across steps while requiring a fresh session for each new task. Inference supports BF16 on NVIDIA GPUs, with optional INT8 and INT4 modes available but with noted performance loss at INT4; example CLIs and a web demo illustrate bounding-box outputs and operation categories.

...Developed by Tongyi-MAI (Alibaba’s research initiative), the MAI-UI models are multimodal agents trained to understand user instructions and corresponding screenshots, grounding those instructions to on-screen elements and generating sequences of GUI actions such as taps, swipes, text input, and system commands. Unlike traditional UI frameworks, MAI-UI emphasizes realistic deployment by supporting agent–user interaction (clarifying ambiguous instructions), integration with external tool APIs using MCP calls, and a device–cloud collaboration mechanism that dynamically routes computation to on-device or cloud models based on task state and privacy constraints.

Qwen-MM-Plugins is a collection of native multimodal capabilities that can extend Qwen-powered agent harnesses beyond text-only workflows. Each capability combines an instructional skill with an optional MCP server that exposes specialized tools. The core package can inspect images, videos, documents, and 3D models while providing OCR, grounding, segmentation, transcription, and visual question answering. Additional modules support long-video memory, audio-video analysis, media generation,...

...The workflow builds an MBB-style evidence table before developing storylines and converging on an SCR narrative. Users select from eight fixed visual systems, after which the skill plans each slide’s hierarchy, grid, charts, palette, and information density. Image-generation blueprints guide reconstruction while native text, shapes, tables, charts, and vectors preserve core editability. Multiple quality gates inspect structure, visuals, overflow, spatial alignment, curves, and editable elements. Delivery includes the PPTX, rendered previews, manifests, and QA results, with failed hard gates blocking completion.