← DeepSeek Harness Plugins
DeepSeek Harness PluginManifest Valid

dsh-v4flash-tiler

Auto-tiles high-resolution chat images into labeled grids to preserve fine visual details for DeepSeek models.

VisionTerminal & TUIDeveloper Tools
1GitHub Stars0ForksUpdated2026-10-24

Install

$ dsh plugin --profile web add github:doublehappy123/dsh-v4flash-tiler

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-v4flash-tiler is a DeepSeek Harness plugin that intercepts oversized chat images and splits them into up to 9 coordinate-labeled sub-tiles with 15% overlap using a local Python tiling engine. This prevents small text, icons, and charts from being degraded by default 800x800 downsampling while giving the vision model explicit grid coordinates to reason across the whole image.

Key Capabilities

  • Automatic pre-step tiling for user images exceeding 1024px on any dimension
  • Preserves fine detail with up to 9 sub-tiles and 15% overlap
  • Provides row and column grid coordinate metadata to the model
  • Includes tiler_route tool to manage tiling behavior across auto, off, and always modes
  • Prevents cross-image tile confusion when multiple images are sent in a single turn
  • Includes standalone CLI and Python API for dry-run inspection and vision analysis

Useful For

  • Reading small text and terminal code in high-resolution screenshots
  • Analyzing dense architectural diagrams, spreadsheets, and charts without downsampling blur
  • Processing complex multi-image chat queries requiring precise coordinate mapping

Who It Fits

  • Developers and researchers relying on DeepSeek vision models for dense image analysis
  • DeepSeek Harness users working frequently with high-DPI screenshots and diagrams

Documented Limitations

  • Requires a separate Python engine installation (pip install ./engine) callable by the host
  • Consumes significantly more context tokens (approximately 384 tokens per tile, up to ~3400 tokens per 9-tile image)
  • Automatic image opening listener and chat_send_image tool are disabled by default and require manual activation
  • CLI manual vision analysis mode requires a configured DEEPSEEK_API_KEY

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
doublehappy123/dsh-v4flash-tiler
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin
Source checked
291a7ed · 2026-10-24

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
18
Last commit
2026-10-24
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