Autonomous coding agent right in your IDE, capable of creating/editing files, running commands, using the browser, and more with your permission every step of the way.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
A fork of Cline that adds a third Image Mode
model channel. When you paste an image into a chat running a non-vision model
(e.g. DeepSeek Reasoner), your configured Image Mode vision model describes the
image, and that description is bridged into the chat as a collapsible, copyable
text block — so non-vision models get full image context without ever receiving
raw image bytes.
Three model channels — Plan, Act, and Image Mode. Each tab in API
Configuration keeps its own provider, model, API key, and reasoning effort.
Image bridge — images are intercepted at send, described by the Image Mode
vision model, and rolled up as a selectable/copyable block.
Capability-aware — the bridge runs only when the active Plan/Act model is
non-vision (or unknown); vision-capable models receive the raw image as usual.
Debug logging — a Settings toggle records each bridge call (provider,
model, URL, image type/size, auth, status) to the output channel and shows the
most recent calls inline under failed bridge blocks, with a one-click toggle
right in the chat.
Setup
Open Settings → API Configuration and pick the Image Mode tab.
Choose a vision-capable provider and model (e.g. DeepSeek, OpenAI, OpenRouter,
or Gemini with a vision model) and enter its API key.
Paste an image into the chat — the bridge describes it and feeds the text
description to your Plan/Act model.
Debug logging
In Settings → API Configuration → Image Mode tab, enable Image bridge debug
logging. Each bridge call is then logged to the output channel (View → Output → Cline Cubed), and the most recent calls are shown inline under the bridge
block. If a bridge call fails, the panel appears automatically with a one-click
way to turn the toggle off.
