Anionex GitHub avatar

codex-deepseek-vision

Anionex

A local proxy that lets text-only DeepSeek models in Codex 'see' images by rewriting pasted images and view_image calls into descriptions from a vision API, without extra MCP, skills, or CLI tools.

Stars

23

7-day growth

No data

Forks

0

Open issues

0

License

No data

Last updated

2026-08-01

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It solves a real integration gap with a small, self-contained proxy: Codex's image tools pass the vision check, the proxy converts images to text descriptions in-flight, and DeepSeek can respond about image content—verified in real Codex + DeepSeek sessions.

Who it is for

  • Codex desktop users who have switched to DeepSeek and need vision support
  • Developers using cc-switch to manage Codex model catalogs
  • Developers who want to add image-to-text rewriting to their own Codex proxy/gateway

Use cases

  • Asking DeepSeek in Codex to describe UI screenshots or design style
  • Debugging interface field mismatches by sending a screenshot to the model
  • Using multiple images in one request with parallel vision description and caching
  • Integrating the rewrite functions into an existing proxy to avoid full installation

Strengths

  • Multi-image requests are processed concurrently, so N images approximate the latency of one image
  • Descriptions are cached by image sha256, avoiding repeated vision API calls for the same image
  • Zero-dependency setup: only a .env with an OpenAI-compatible vision API key is required
  • Handles both pasted image message.content and view_image function_call_output.output structures

Considerations

  • This is text-description vision, not true multimodal input; quality depends on the configured vision model
  • Requires a valid vision API key or local CLI fallback; otherwise DeepSeek still replies it cannot see the image
  • Verified on macOS with the Codex desktop app; custom gateways must manually merge the rewrite logic and catalog image modality

README quick start

快速开始

Description

一个可以让像deepseek-v4这样的纯文本模型, 在codex中无障碍调用内置看图工具(view_image)的方案

Related repositories

Similar projects matched by category, topics, and programming language.

lopopolo
Featured
lopopolo GitHub avatar

harness-engineering

Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.

AI & Machine LearningAI Agents
2,390
slvDev
Featured
slvDev GitHub avatar

esp32-ai

A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.

AI & Machine LearningLarge Language Models
1,960
littledivy
Featured
littledivy GitHub avatar

mimic

mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.

AI & Machine Learning
1,482