- unwind ai
- Posts
- Run Claude Code from Anywhere
Run Claude Code from Anywhere
PLUS: Claude Sonnet 4 with 1M context window, OpenAI won gold at international programming competition
Shubham Saboo & Gargi Gupta
August 13, 2025
Today’s top AI Highlights:
Opensource AI router for MCP servers and LLMs
Run Claude Code from anywhere - terminal, web, or mobile
Claude Sonnet 4 now supports 1M tokens in the API
After IMO 2025, OpenAI just won gold at IOI 2025
Just enter your domain, and this opensource AI agent will hack it
& so much more!
Read time: 3 mins
AI Tutorial
Our Awesome LLM Apps repo has over 100+ AI Agents and RAG apps.
We’re giving away an insane AI workflow for free that lets you build your own AI Agent in under 3 minutes with zero coding experience.
Here's the exact 3-step process:
Step 1: Find the Blueprint
↳ Browse the Awesome LLM Apps repository on GitHub
↳ Pick any AI agent that solves your problem
↳ Copy the entire repo URL
Step 2: Create the Super Prompt
↳ Drop the GitHub URL into gitingest
↳ Get an LLM-friendly version of the entire codebase
↳ Copy the generated prompt
Step 3: Let AI Build It
↳ Paste the prompt into Gemini (preferred because of long context)
↳ Ask for your custom version
↳ Get working code, README, and requirements in minutes
By following this workflow, you can create a fully functional AI agents that would have taken days to code manually. Watch the full tutorial here 👇
We share hands-on tutorials like this every week, designed to help you stay ahead in the world of AI. If you're serious about leveling up your AI skills and staying ahead of the curve, subscribe now and be the first to access our latest tutorials.
Latest Developments
Your AI agents are juggling multiple MCP servers and LLM providers like a chaotic orchestra - it’ll eventually break.
Nexus is an opensource AI router that optimizes how AI agents interact with multiple MCP servers and LLM providers. Built in Rust for maximum performance, Nexus consolidates all your MCP servers into a single unified interface. It also automatically selects the best model for each task based on performance, cost, and availability requirements.
The router eliminates the complexity of managing point-to-point connections between your AI agents and external services, replacing dozens of individual integrations with one streamlined endpoint.
Key Highlights:
Unified Architecture - Consolidates multiple MCP servers and LLM providers into a single endpoint, eliminating complex point-to-point connections and reducing system maintenance overhead.
Tool Discovery - Exposes only two tools to your AI agents ('search' and 'execute') instead of overwhelming them with hundreds of individual MCP tools, using natural language discovery and automatic namespacing.
LLM Routing - Automatically routes requests to optimal language models based on task type, latency requirements, context length, and cost considerations with built-in failover support.
Observability - Monitor and analyze the performance of your models in real-time, identify bottlenecks, and optimize your routing strategy.
Your AI agents are busy coding while you're grabbing coffee, but what happens when they hit a roadblock and you're nowhere near your terminal?
You come back hours later to find them stuck on a simple question that could have been answered in seconds.
Omnara solves this by turning your phone into mission control for Claude Code and other AI agents, letting you monitor their progress and respond to their questions from anywhere. It’s an open-source "agent command center" that provides real-time visibility into what your agents are doing, complete with push notifications when they need your input.
The platform supports seamless switching between terminal, web, and mobile interfaces, maintaining the full native Claude Code experience across all devices.
Key Highlights:
Device switching - Start a Claude Code session in your terminal and continue responding to the same conversation from your phone or web dashboard without losing any context or functionality.
Push notifications - Get push alerts only when your agents actually need human input, eliminating the frustration of returning to find stuck or failed jobs hours later.
AI agent support - While launching with Claude Code integration, Omnara's framework works with any AI agent that needs human-in-the-loop capabilities, from custom bots to workflow automation tools.
Opensource - The entire backend is available under Apache 2.0 license on GitHub. The web and mobile app also have a free tier, offering up to 10 agent sessions per month before the $9/month unlimited plan.
Used by Execs at Google and OpenAI
Join 400,000+ professionals who rely on The AI Report to work smarter with AI.
Delivered daily, it breaks down tools, prompts, and real use cases—so you can implement AI without wasting time.
If they’re reading it, why aren’t you?
Quick Bites
Claude Sonnet 4 now supports up to 1 million tokens of context on the Anthropic API in beta - a 5x increase that lets you process entire codebases with over 75,000 lines of code or dozens of research papers in a single request. Requests exceeding 200K tokens are automatically charged at premium rates (2x input, 1.5x output pricing).
After the International Math Olympiad 2025, OpenAI’s “reasoning system” scored high enough to achieve gold in one of the world’s top programming competitions - the 2025 International Olympiad in Informatics (IOI), placing them #6 when ranked with humans and #1 when ranked with other AIs. The model operated under identical constraints: 5-hour time limits, 50 submission caps, and no internet access. This is a dramatic leap from 49th percentile performance just one year ago.
Local AI search has now caught up to the cloud. Menlo Research released Jan-v1, a 4B parameter model that beats Perplexity Pro's accuracy while running entirely on your local machine. Built on Qwen’s latest 4B-Thinking model, the model achieves 91% SimpleQA accuracy and handles web search plus deep research tasks. Available now on Hugging Face in both standard and GGUF formats for llama.cpp integration.
Perplexity has offered an audacious, unsolicited $34.5 billion to buy Google Chrome in an all-cash deal. That’s more than twice Perplexity’s own estimated valuation of $14 billion. This comes after the DOJ alleged that Google is creating a monopoly in search, and proposed to sell Chrome. However, Google vowed not to sell the browser. If this is Perplexity’s negotiation strategy, they might as well throw in an offer for the moon while they’re at it.
Genspark's new AI meeting notetaker turns your Apple Watch into a meeting assistant, capturing audio and generating summaries with just one prompt. It syncs effortlessly with your calendar and can highlight key discussions, topics, and action items for easy sharing. Integrates with AI tools like Slides, Sheets, and Docs for easily creating beautiful professional assets.
Tools of the Trade
Strix: AI agent that interacts with your application like a real attacker would to discover code vulnerabilities. It integrates into development workflows to provide real-time security assessment of codebases, repositories, and web applications with automated patching and detailed reports.
Meteor: Chromium-based agentic browser that can handle actions on your behalf. Integrates AI agents directly into a browser interface to handle routine web tasks such as data entry, calendar management, and multi-site research.
Dereference: Run multiple AI model sessions (Claude, GPT-5, Gemini) simultaneously with the ability to branch conversations at any point. It provides Git-like conversation management where you can explore different solution paths and merge successful branches back into the main conversation flow.
Awesome LLM Apps: Build awesome LLM apps with RAG, AI agents, MCP, and more to interact with data sources like GitHub, Gmail, PDFs, and YouTube videos, and automate complex work.
Hot Takes
gpt-5 hate is 80% skill issue ~
mephistoOpenAI's decision to give their Codex CLI tool and their Codex asynchronous coding agent the exact same name means that when people say "have you tried Codex?" i have genuinely no idea which one they are talking about ~
Simon Willison
That’s all for today! See you tomorrow with more such AI-filled content.
Don’t forget to share this newsletter on your social channels and tag Unwind AI to support us!
PS: We curate this AI newsletter every day for FREE, your support is what keeps us going. If you find value in what you read, share it with at least one, two (or 20) of your friends 😉
Reply