← DeepSeek Harness Plugins
DeepSeek Harness PluginManifest Valid

dsh-ollama-vision-bridge

DeepSeek Harness bridge plugin that automatically generates local Ollama VL descriptions for text-only models.

VisionTerminal & TUIDeveloper ToolsRemote Execution
1GitHub Stars0ForksUpdated2026-10-19

Install

$ dsh plugin --profile web add dsh-ollama-vision-bridge

Plugin Overview & Capabilities

AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.

source-grounded
dsh-ollama-vision-bridge is a DeepSeek Harness plugin that intercepts image uploads directed to text-only chat models. Instead of triggering model rejection, it leverages a local Ollama vision-language model to describe the image and injects the description into the prompt context in the same execution step, complete with keep_alive VRAM cooling.

Key Capabilities

  • Automatic prompt interception and local Ollama VL image description generation for text-only models
  • Single-step context injection via agent waterfall hook without modifying user message history
  • Configurable Ollama keep_alive timing for automatic VRAM unloading and cooling
  • Built-in CLI commands to setup, switch VL models, check service status, and test vision endpoints
  • Idempotent patching of dsh-api-session-controller and profile-based Cordis patch integration

Useful For

  • Uploading diagrams, screenshots, or photos to text-only LLMs without incurring cloud vision API fees
  • Running local privacy-focused image analysis using local Ollama models like qwen3-vl
  • Conserving GPU VRAM by automatically offloading local vision models after idle timeouts

Who It Fits

  • DeepSeek Harness users running local Ollama instances
  • Developers wanting multimodal workflows with text-only foundation models
  • Privacy-conscious users who prefer local image inference

Documented Limitations

  • Requires a running local Ollama service with an installed vision-language model
  • Relies on code patching of dsh-api-session-controller which must be reapplied after host updates
  • Only covers primary Web GUI chat sessions; parallel subagent entry points are not bridged
  • Native VL chat models bypass the bridge directly

DSH Compatibility

Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.

Not tested yetNo runtime compatibility tests have been published yet.

Security Signals

Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.

Dependency Counts

package.json declares 0 runtime, 0 development, 1 peer, and 0 optional dependencies.

info
Dsh Bundle Declared

package.json declares DSH bundle metadata.

info
License Declared

GitHub reports the repository license as MIT.

info
Package Manifest Available

A root package.json was captured and can be inspected by the registry.

info
Source Available

Public GitHub source metadata is available for this registry snapshot.

info

Source & Registry Notes

Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.

Source repository
nvbb/dsh-ollama-vision-bridge
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin
Source checked
440deae · 2026-10-19

This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.

Repository Activity

GitHub stars
1GitHub stars
Forks
0
Open issues
0
Last commit
2026-10-19
Last release
No release detected
For maintainers

Maintaining this plugin?

This listing is generated from public repository data. If you maintain this project, you can review the information and share this listing with your users if you find it useful.

Add to README
Listed on DSHPlugin.app