Trending on GitHub — OffNet Newsroom — Friday, September 25, 2026

GitHub Trending (daily) githubrepos ⚠ unverified date/source

Google Ax: Declarative Agentic Orchestration Runtime for Scale

Google released Ax, an open-source runtime designed to orchestrate autonomous agent workloads at scale. It uses a Kubernetes-like declarative model to define tasks, workspaces, and model specifications, handling sandboxing and execution via Agent Substrate. The project currently targets high-throughput clusters capable of billions of tasks but warns that core protocols are still evolving.

Ax uses declarative YAML (task.yaml) to define agent tasks, workspaces, and model specs.
Leverages Agent Substrate for sandboxed execution, ensuring isolation for autonomous agents.
Architecture mirrors Kubernetes concepts, making it intuitive for infrastructure engineers.
Project is pre-stable with active refinement; expect breaking changes before release.
Designed for high-throughput clusters, aiming to run billions of tasks per cluster.

Read at source →

GitHub Trending (daily) githubrepos ⚠ unverified date/source

Hindsight: Agent memory system focused on long-term learning over simple recall

Vectorize-io's Hindsight is a memory architecture designed for AI agents to learn over time rather than merely retrieving conversation history. The system positions itself as a superior alternative to standard RAG and knowledge graph approaches for long-term memory tasks. It claims state-of-the-art performance by implementing mechanisms for retaining, recalling, and reflecting on past interactions.

Replaces static RAG with dynamic learning capabilities for persistent agent memory.
Supports LLM wrappers, coding agents, and MCP integration for flexible deployment.
Claims better long-term memory accuracy than traditional knowledge graph methods.
Available as a cloud service, server, or embedded component for various platforms.

Read at source →

GitHub Trending (daily) githubrepos ⚠ unverified date/source

Univer SDK: Open-source Office suite for embedding in AI agent apps

Univer is an open-source SDK that provides a unified runtime for building spreadsheet, document, presentation, and relational table experiences within custom applications. It features a plugin architecture, canvas-based rendering, and a formula engine, exposing a single Facade API compatible with both browser and Node.js environments. The project targets developers needing to embed productivity features directly into their products without relying on external office suites.

Provides a unified API for spreadsheets, docs, slides, and tables in one SDK.
Runs in browser and Node.js via a single Facade API surface.
Supports plugin architecture and canvas-based rendering for customization.
Includes a built-in formula engine for spreadsheet capabilities.
PDF support is listed as coming soon, not currently available.

Read at source →

GitHub Trending (daily) githubrepos ⚠ unverified date/source

obra/superpowers introduces agentic skills framework for coding agents

Superpowers is a software development methodology designed for coding agents, built on composable skills and initial instructions. It aims to standardize how agents like Claude Code, Cursor, and Devin approach coding tasks. The framework provides a structured workflow to handle both basic tasks and error recovery scenarios.

Provides a standardized methodology for various AI coding agents rather than just a tool.
Uses composable skills to structure agent behavior from initialization to error handling.
Supports major platforms including Cursor, Devin, GitHub Copilot, and Claude Code.
Focuses on reliable agent workflows through defined skills and instruction sets.

Read at source →

GitHub Trending (daily) githubrepos ⚠ unverified date/source

Treg: OpenRouter-style abstraction for AI agent tools

Treg acts as a unified interface for over 3,000 agent tools across 60+ providers, similar to how OpenRouter handles LLMs. It allows agents to execute tasks like SEO, scraping, or media generation via a single base URL and token, removing the need for individual provider signups. The system supports private team keys and CLIs, keeping credentials server-side while enabling agent access.

Centralizes 3,000+ tool endpoints behind one API, simplifying agent orchestration.
Eliminates per-provider account management and credential distribution to agents.
Pricing is per-call, starting from one cent, with transparent cost visibility.
Supports internal team keys and CLIs for secure, shared agent capabilities.
Covers diverse use cases: SEO, backlinks, enrichment, ads, and media generation.

Read at source →

GitHub Trending (daily) githubrepos ⚠ unverified date/source

stable-diffusion.cpp adds Day-0 support for Qwen-Image-2.1 and LTX-2.5

The open-source C/C++ inference engine for diffusion models has expanded its capabilities with immediate support for Qwen-Image-2.1 and LTX-2.5. The project also recently integrated LTX-2.3, Krea2, Ideogram4, and other emerging architectures. Maintained as a high-performance, dependency-light alternative, it continues to track new model releases with rapid turnaround times.

Day-0 support for Qwen-Image-2.1 enables immediate experimentation with Qwen's latest image generation capabilities.
LTX-2.5 integration extends the engine's reach into newer video and image diffusion architectures.
Pure C/C++ implementation offers a low-overhead, portable inference layer for edge or embedded deployments.
Frequent API updates mean developers should pin versions or monitor breaking changes closely.
Support for diverse models like MiniMax-H3 and Krea2 positions this as a unified inference backend.

Read at source →