⚡ earendil-works/pi is making waves. Here's the full picture.

🔗 https://github.com/earendil-works/pi
📝 AI agent toolkit: coding agent CLI, unified LLM API, TUI & web UI libraries, Slack bot, vLLM pods
──────────────────────────────

The pi agent harness project is a self-extensible coding agent that combines a range of tools to enhance development workflows. At its core, the project includes three key packages: @earendil-works/pi-coding-agent for interactive coding, @earendil-works/pi-agent-core for agent runtime and state management, and @earendil-works/pi-ai for unified multi-provider large language model (LLM) APIs.

To get started, users can visit the project website for demos, read the documentation, or ask the agent to explain itself. The project encourages users to share their open-source coding agent sessions to improve the agents with real-world tasks and tool use.

From a technical perspective, the project is built using a range of technologies and includes features like differential rendering and a terminal UI library. The project also prioritizes security and supply-chain hardening through measures like pinning direct external dependencies and verifying pinned direct dependencies.

Audience: The project is geared towards developers and coding enthusiasts looking to enhance their workflows and explore the potential of self-extensible coding agents.

One-liner takeaway: Supercharge your coding workflow with the pi agent harness project, a cutting-edge self-extensible coding agent that's open-source and constantly evolving.

──────────────────────────────
🧠 Channel: https://t.me/GithubRe

🔍 Deep-diving into Alishahryar1/free-claude-code — fresh off the trending list.

🔗 https://github.com/Alishahryar1/free-claude-code
📝 Use claude-code for free in the terminal, VSCode extension or discord like OpenClaw (voice supported)
──────────────────────────────

The Free Claude Code repository is a drop-in proxy for Claude Code's Anthropic API calls, allowing users to choose from 17 different provider backends, including NVIDIA NIM, OpenRouter, and Google AI Studio. This proxy enables per-model routing, streaming, and local request optimizations, making it a versatile tool for developers.

The key features of this repository include:
- Seventeen provider backends to choose from, each with its own unique models and capabilities
- Per-model routing, allowing users to send different types of traffic to different providers
- Native Claude Code model picker support, making it easy to switch between models
- Optional integrations with Discord, Telegram, and VS Code, providing a range of use cases

To get started, users can follow the Quick Start guide, which includes installing the required dependencies, starting the proxy server, and configuring the proxy settings through the Admin UI.

From a technical standpoint, the repository uses

Python 3.14

and is tested with Pytest, ensuring that the code is reliable and stable. The repository also includes type checking with Ty and code formatting with Ruff, making it easy to contribute to and maintain.

The target audience for this repository includes developers who want to use Claude Code with different provider backends, as well as those who want to customize their AI workflows.

In summary, the Free Claude Code repository provides a flexible and customizable solution for working with Claude Code and Anthropic API calls, making it an ideal choice for developers who want to take control of their AI workflows. With its wide range of features and integrations, it's an essential tool for anyone looking to unlock the full potential of AI - and with this proxy, the possibilities are endless.

──────────────────────────────
🧠 Channel: https://t.me/GithubRe

🔥 colbymchenry/codegraph is trending — and it deserves your attention.

🔗 https://github.com/colbymchenry/codegraph
📝 Pre-indexed code knowledge graph for Claude Code, Codex, Cursor, OpenCode, and Hermes Agent — fewer tokens, fewer tool calls, 100% local
──────────────────────────────

CodeGraph is a powerful tool that supercharges Claude Code, Cursor, Codex CLI, opencode, and Hermes Agent with semantic code intelligence. It provides a pre-indexed knowledge graph of symbol relationships, call graphs, and code structure, allowing agents to query the graph instantly instead of scanning files. Key features include smart context building, full-text search, impact analysis, and framework-aware routes. codegraph is 100% local, with no data leaving your machine, and supports 19+ languages.

To get started, simply run the installer with npx @colbymchenry/codegraph, then restart your agent and initialize your project with codegraph init -i. With CodeGraph, you can enjoy average savings of 35% cheaper, 57% fewer tokens, 46% faster, and 71% fewer tool calls. One-liner takeaway: CodeGraph is a game-changer for coding efficiency, and it's just one install away.

──────────────────────────────
🧠 Channel: https://t.me/GithubRe

⚡ multica-ai/multica is making waves. Here's the full picture.

🔗 https://github.com/multica-ai/multica
📝 The open-source managed agents platform. Turn coding agents into real teammates — assign tasks, track progress, compound skills.
──────────────────────────────

Multica is an open-source managed agents platform that turns coding agents into real teammates. You can assign tasks to agents like you'd assign to a colleague, and they'll pick up the work, write code, report blockers, and update statuses autonomously. Key features include agents as teammates, squads for grouping agents and humans, autonomous execution, autopilots for scheduling recurring work, reusable skills, unified runtimes, and multi-workspace support.

To get started, you can install Multica using a brew install command or an install script. Then, set up and start the daemon using the multica setup command. Verify your runtime, create an agent, and assign your first task.

The platform is designed for human + AI teams and supports various agent CLIs, including Claude Code, Codex, and GitHub Copilot CLI. The architecture consists of a Next.js frontend, a Go backend, and a PostgreSQL database.

Developers can contribute to the Multica codebase by following the CONTRIBUTING.md guide, which includes prerequisites, development workflow, and troubleshooting. With Multica, a small team can move like a large one - your next 10 hires won't be human.

──────────────────────────────
🧠 Channel: https://t.me/GithubRe

🔥 shiyu-coder/Kronos is trending — and it deserves your attention.

🔗 https://github.com/shiyu-coder/Kronos
📝 Kronos: A Foundation Model for the Language of Financial Markets
──────────────────────────────

Kronos is a foundation model for financial markets, specifically designed to handle the unique characteristics of financial data. It's a decoder-only model, pre-trained on data from over 45 global exchanges. The model uses a two-stage framework, consisting of a specialized tokenizer that quantizes continuous, multi-dimensional K-line data into hierarchical discrete tokens, and a large, autoregressive Transformer that is pre-trained on these tokens.

The model is open-source and readily accessible from the Hugging Face Hub. A live demo is available to visualize Kronos's forecasting results.

To get started, simply install the dependencies, load a pre-trained model and its corresponding tokenizer, and instantiate the predictor. The predictor can be used to generate forecasts for given input data.

Technical highlights include a novel two-stage framework, a specialized tokenizer, and a large, autoregressive Transformer. The model is designed for quantitative tasks and can be fine-tuned for specific use cases.

Kronos is suitable for data scientists and quantitative researchers who work with financial data.

In a nutshell, Kronos is a powerful tool for financial forecasting, and its open-source nature makes it accessible to everyone: forecast your financial future with Kronos today!

──────────────────────────────
🧠 Channel: https://t.me/GithubRe

🎯 manaflow-ai/cmux landed on trending. Worth a proper look.

🔗 https://github.com/manaflow-ai/cmux
📝 Ghostty-based macOS terminal with vertical tabs and notifications for AI coding agents
──────────────────────────────

cmux is a Ghostty-based macOS terminal that streamlines your coding workflow with features like vertical tabs, notifications, and an in-app browser. Key features include notification rings, a notification panel, in-app browser, vertical and horizontal tabs, SSH support, and Claude Code Teams integration.

To get started with cmux, you can download the DMG file from the GitHub releases page or install it using Homebrew with brew tap manaflow-ai/cmux and brew install --cask cmux.

Technical highlights of cmux include its native macOS app built with Swift and AppKit, GPU-accelerated rendering, and scriptable CLI and socket API. cmux is designed for developers who want a customizable and efficient coding workflow, and its target audience includes anyone looking for a powerful terminal replacement.

cmux is a primitive, not a solution - it provides a set of building blocks for you to create your own workflow. With its cmux.json configuration file, you can define custom commands and automate your workflow.

In short, cmux is a game-changer for coding workflows - it's not just a terminal, it's a launchpad for your coding productivity.

──────────────────────────────
🧠 Channel: https://t.me/GithubRe