A self-hosted drop-in replacement for TypeSafe's jev, powered by GliFormer.
Calibrated 151M Non-Autoregressive Decision Engine beating TypeSafe Jev & Laya on LocalLLaMA/typed-decisions (77.10% acc, 0.0636 Brier, 0.0144 ECE)
A self-hosted coding agent for background work - works with Claude Code, Codex, OpenCode & 100+ providers
The open-source design agent and harness, better than Claude Design on academic communication artifacts production. This DesignHarness can also be used with any coding harness you like ( Codex/Claude Code/Kimi Code/Pi/OpenCode etc..) and any agentic model you want.
Encoder-Free Unified Image and Video Understanding and Generation
One command turns a DGX Spark or ASUS Ascent GX10 into a private AI server: seven measured, pinned AI lanes (text up to 1M tokens, images, video with sound), run from a local cockpit on a laptop or a phone. Nothing leaves the box.
Swap a large text encoder for a small one with a learned linear projection. MiniMax H3 conditioning from 15.7 GB down to 5.2 GB.
Hide, encode, and decode encrypted files, messages, and alternate audio streams in music hosted on popular streaming platforms (Spotify, Tidal, Bandcamp, etc).
Migrate coding agent sessions across 18 harnesses (Claude Code, Codex, Pi, OpenCode, Copilot, Antigravity, Vibe, Muse, Qwen, Kimi, Cursor, etc.)
Your durable workspace across AI agents.
Generalist Multitask Transformer Encoders
Agent Skill: realistic handwriting with the Inko API — solve & hand-write homework, write on your own notebook paper, phone-photo looks, PDF. For Claude Code, Codex, Cursor, Copilot, Gemini CLI, OpenCode.
WeeLLM runs large diffusion models with as little as 4 GB of VRAM, without any quantization. It dynamically determines how many layers can fit within the available VRAM and streams the text encoder and transformer layers to the GPU layer by layer, enabling inference on hardware with limited VRAM. It supports both safetensors and GGUF models.
Build an AI Coding Agent from Scratch in Pure Python, Step by Step | 用纯 Python 从零搭建 AI 编程 Agent,边实现边理解其工作原理
A unified audio-video variational autoencoder with cross-modal alignment for joint reconstruction and generation
An Agent Skill for Android phone control via ADB: observe, act, and verify. For Pi, Codex, Claude Code, and OpenCode.
A local acoustic control system that learns user-defined labels for the surface or location of knocks, such as desk or wall. It also counts threshold-qualified knocks within each captured window. By default, a frozen Microsoft BEATs Iter3+ AS2M encoder creates 768-D audio embeddings and an incremental classifier learns the user's labels.
Native macOS workspace with one local knowledge graph across Claude Code, Codex, OpenCode, and Cursor.
Universal Zero-Refusal Agent Framework - 40+ jailbreak templates (T1-T119), 1128 curated skills, executable framework-core, installer untuk Hermes/Claude/Codex/opencode. Verified.
A local OpenAI-compatible bridge to the free DeepSeek web chat, which you can plug into opencode as a regular model provider and use as a full-fledged agent (reading/editing files, bash, code search, etc.) — with no API key and no payment.
Claude Skills that encode chemical process and manufacturing engineering practice
Profanity-triggered self-audit protocol for coding agents (Claude Code, Codex CLI, Kimi CLI, OpenCode). Python stdlib only.
An agentic playbook for pre-filling the Bangladesh NBR e-Return (etaxnbr.gov.bd) using Opencode (but anything will work) + Playwright MCP: an AI agent walks the return wizard, patches your draft against your own data files, and stops at Save draft — you review, pay, and submit yourself.
Kunpeng-Skill is a powerful multimodal distillation toolkit that transforms high-value insights from repositories, websites, UIs, videos, images, and documents into reusable methodologies, model-agnostic regeneration specifications, and more.
Runs opencode through rotating Tor exit lanes so free-tier rate limits and stalls stay invisible -- with live per-request proof of the lane and exit IP behind every call.
Procedural writing-style skills for AI agents: official frameworks (plain language, business writing, GOV.UK, AP style, STE-100, developer docs) + character voices. Agent Plugins 1.0.0, works in opencode, Claude Code, Codex.
Curated developer tools, AST analyzers, and automated verification workflows.
Automated data preparation for machine learning. Load a dataset, clean, transform, encode, scale and split it, explore it with built-in analysis, and export model-ready data with a full report. Built with Python and Streamlit.
Pragmatic multi-agent orchestration: independent review and steerable delegation across Codex, Claude Code, OpenCode, Grok Build, Gemini, and Devin CLI — porch CLI + Agent Skills
OpenCode Go 用量统计面板 · Local-first usage dashboard (Python + pywebview)
Phoneme alignment based on a variational autoencoder (VAE) and self-supervised learning (SSL) acoustic features.
Optimized local serving engine for Kimi-Linear-48B: INT4 quantizer, fused decode kernels for a measured 3.18x, and an OpenAI-compatible server. Ships with k3, a bridge that detects the client per request so Claude Code, Codex, Cline, Aider and opencode all work unchanged.
Perhaps the only agent skill you need: if memory loss is the root, let every AI agent rebirth follow the First-Time Principle. That's all.
FASTDIAR: Frame-Level Speaker Encoder for Streaming Diarization
Open visual probability research: encode an image once, answer multiple structured questions. Public data, trained weights, reproducible baselines and failure cases.
Long-term memory for AI coding agents (Claude Code, Codex, opencode) as a tree of plain files
OpenCode Go ????? (5h/?/? ????, ?????????, free??????)
Use Claude Design from Claude Code, Codex, Cursor, Gemini CLI, and 70+ coding agents. Create designs from your real codebase, refine them on the canvas, and sync every change back. One installer, no MCP setup.
A reusable opencode skill that scaffolds a complete team of AI agents in any project.
A scheduled, agent-driven software factory. Keep a plain JSON backlog; Forgeo runs your coding agent (Claude Code, Codex, opencode) on each task and commits to main. Refactors when idle. No branches, no PRs.
CLI for discovering local skills and agent profiles and installing them across coding clients through format adapters.
Unofficial PoC: a BERT-style encoder decision engine behind a typed-decision (noul / choice / score) HTTP API. Not affiliated with TypeSafe.
Official code release for "MoE-ViE Mixture of Experts Vision Encoder for Efficient Image and Video Understanding" (ECCV 2026)
LightVAE: Fast video VAE decoders for Wan2.1, Wan2.2, and MiniMax-H3, plus a compressed H3 reference encoder. Pretrained weights and ComfyUI workflows.
CLI-инструмент для поиска вакансий, откликов и поднятия резюме на hh.ru через LLM агентов: Claude Code, Cursor, OpenCode и др.
Safely reclaim disk space from bloated OpenCode SQLite databases
Full-spectrum security research workbench for opencode — 284 MCP tools, 207 KB articles, Ghost security-operator mode, max reasoning by default
Automated Cloudflare WARP IP rotator and proxy server for OpenCode Zen. Eliminates HTTP 429 rate limits, persists usage metrics in SQLite, and provides a clean management dashboard.
RDV AIR – ИИ-агент для двухэтапного ревью кода 1С с ролями аудитора и критика, MCP и запуском из Jira
Structured decisions over Russian text in one encoder pass: choice, scale, yes/no, ranking
Streaming ASR, audio captioning and speech QA over a whisper-style encoder with pluggable LLM backends
本地反向代理:在 Trae 与 OpenCode Go (Zen API) 之间做双向转发,自动注入上游要求的会话/请求头(`x-opencode-session` 等),并做模型名别名重写。解决 Trae 直连时报 `Request is missing x-opencode-session` 400 错误的问题。
Lower AI refusal rates in OpenCode: code-context scaffolding, multi-stage workflow, auto refusal repair, local Web UI, zero deps. OpenCode 破甲:代码上下文补全 + 被拒自动修复 + 本地 Web UI,零依赖。
Encode, decode and plot audio with 10 neural audio codecs and audio VAEs (DAC, EnCodec, SNAC, Stable Audio Open, Music2Latent, ACE-Step and more), all through one interface. Useful for looking inside each model's latent space and comparing them side by side.
Reading diffusion language models for decisions: same OpenJev model, one pass instead of 3.65, +5 points, calibrated, no training. Plus cheap task branches on a frozen encoder.
official implementation of CoVisco: Codec-Native Vision Encoder with Native Token Compression for Efficient Unified Visual Understanding
Official implementation of "GRACE: Generation-Aware Latent Compression for Efficient Video Generation" (arXiv 2610.10524)
solving nethack together! @:@:@:@:@
OpenCode Free Tier Fix for Hermes Agent
[NIPS 2026🔥] DC-SAE: Deep Compression Semantic Autoencoder for Faster Diffusion Convergence