Back to directory
Docs & Writingtomoyo123/pi-vision-bridge-tomoyo
pi-vision-bridge-tomoyo
Vision bridge for text-only models (e.g. DeepSeek V4 Flash): images pasted into the chat and screenshots are described by a cheap vision model (MiniMax M3 via opencode.go) and fed to the main model as text. Registers a describe_image tool for on-demand sc
Suggested install command
npm view pi-vision-bridge-tomoyoAlways inspect the linked repository and skill instructions before running commands. Skills are instructions; permissions and execution still matter.
Compatibility
Agent support matrix
3 supported
| Agent | Status |
|---|---|
| Claude Code | Supported |
| OpenCode | Supported |
| Cursor | Not listed |
| MCP | Not listed |
| GitHub Copilot | Not listed |
| Windsurf |