Releases·Tool Release
official-llm-zhazhiji: CLI harness drives seven official web LLMs (DeepSeek, Gemini, Grok, more) as coding-agent brains
GitHub project official-llm-zhazhiji (official LLM juicer) ships separate CLI harnesses that drive official web LLM chatbots—not APIs—as external brains for local coding agents. Seven independent repos cover DeepSeek, Doubao, Gemini, Grok, Qwen, ChatGLM, and Metaso, each with its own CLI and persistent login handling. The project documents per-site rate limits and captcha risk, positioning usage as low-frequency and within each site's policy.
29 engagement1 sourceFri, Sep 11, 2026, 07:45 AM
AI Overview
- GitHub project official-llm-zhazhiji drives official web LLMs (not APIs) via per-brain CLIs — dsb, dbb, gmb, grb, qwb, cgb, msb — turning them into external brains for coding agents
- Seven independent brain repos include DeepSeek, Doubao, Gemini, Grok, Qwen, ChatGLM, Metaso; project says seven is current scale, not a fixed cap
- Risk-control observations: ChatGLM prompts CAPTCHA from the 3rd rapid chat (n=8); Qwen silently throttles ~6 image generations/hour; DeepSeek, Gemini, Grok log zero triggers
- Capabilities differ by brain: Gemini generates 2816×1536 images, Metaso deep research writes 10,000+ word cited reports, Qwen allows anonymous Q&A
- Model selection varies: Gemini offers Flash-Lite/Flash/Pro, Grok Fast/Build free with Auto/Expert/Heavy subscription-gated, Doubao quick/2.1 Turbo
Entities
Related
Releases·Tool Release
llm-openrouter 0.7 adds LLM 0.32 compatibility and server-side tools
Simon Willison released llm-openrouter 0.7, updating the plugin for compatibility with LLM 0.32 and improving support for reasoning models via OpenRouter's Responses API. The release adds three new server-side tools: Shell, WebFetch, and likely others, expanding the plugin's capabilities for LLM users.
Simon Willison released llm-openrouter 0.7, updating the plugin for compatibility with LLM 0.32 and improving support for reasoning models via OpenRouter's Responses API. The release adds three new server-side tools: Shell, WebFetch, and likely others, expanding the plugin's capabilities for LLM users.
0 engagement1 sourcerssFri, Aug 21, 2026, 04:58 PM
Releases·Tool Release
FreeDeepseekAPI-EN: Atharvotech maintains OpenAI-compatible local proxy for DeepSeek Web Chat
Atharvotech maintains FreeDeepseekAPI-EN, an English-adapted version of a local OpenAI-compatible API proxy for DeepSeek Web Chat. The proxy lets users connect DeepSeek to Open WebUI, Claude Code, and OpenAI SDK tools without an official API. Core logic belongs to original author ForgetMeAI.
Atharvotech maintains FreeDeepseekAPI-EN, an English-adapted version of a local OpenAI-compatible API proxy for DeepSeek Web Chat. The proxy lets users connect DeepSeek to Open WebUI, Claude Code, and OpenAI SDK tools without an official API. Core logic belongs to original author ForgetMeAI.
25 engagement1 sourcegithubSun, Sep 6, 2026, 05:07 AM
Releases·Tool Release
llm-eyes tool lets any blind LLM see via tiny local VLM
A GitHub project called llm-eyes bolts a tiny vision-language model (Qwen3.5-0.8B) onto any machine as a local vision service, enabling text-only models like DeepSeek V4 Flash, Qwen, and Llama to caption images, watch webcams, or read video. It runs on Mac (MLX), PC+NVIDIA (llama.cpp), and DGX Spark, and exposes an OpenAI-compatible API.
A GitHub project called llm-eyes bolts a tiny vision-language model (Qwen3.5-0.8B) onto any machine as a local vision service, enabling text-only models like DeepSeek V4 Flash, Qwen, and Llama to caption images, watch webcams, or read video. It runs on Mac (MLX), PC+NVIDIA (llama.cpp), and DGX Spark, and exposes an OpenAI-compatible API.
23 engagement1 sourcegithubTue, Jul 7, 2026, 04:52 PM
Products·launched
Browser-based LLM inference and model manager using WASM (wllama)
Goku is a web application that runs large language models directly in the browser using WebAssembly (via wllama). It allows users to download, manage, and run LLMs locally without server-side infrastructure, solving privacy and latency issues for on-device AI inference.
Goku is a web application that runs large language models directly in the browser using WebAssembly (via wllama). It allows users to download, manage, and run LLMs locally without server-side infrastructure, solving privacy and latency issues for on-device AI inference.
2 engagement1 sourcehackernewsWed, Jul 15, 2026, 01:36 PM
Products·launched
72 single-purpose CLI tools powered by local Ollama models
A collection of 72 small command-line tools that each perform a single task (e.g., writing commit messages, explaining commands, translating English to shell commands, debugging errors) using a local Ollama model. They read from stdin and write to stdout, enabling pipeable, offline LLM-assisted workflows in the terminal without any API key.
A collection of 72 small command-line tools that each perform a single task (e.g., writing commit messages, explaining commands, translating English to shell commands, debugging errors) using a local Ollama model. They read from stdin and write to stdout, enabling pipeable, offline LLM-assisted workflows in the terminal without any API key.
86 engagement1 sourceredditTue, Jul 14, 2026, 01:04 PM