Pulse

A new web application called Ithihāsas has been released, designed to explore characters and relationships within Hindu epics like the Ramayana and Mahabharata. The tool utilizes interactive visualizations such as force graphs, dynasty trees, and chord diagrams to present complex narrative structures. Ithihāsas aims to offer a modern lens through which users can engage with ancient wisdom. AI

Mcptube-vision is a new tool that transforms YouTube videos into a persistent knowledge base by analyzing both transcripts and visual frames. It builds upon the concept of a "Karpathy LLM Wiki," where knowledge compounds with each added video, unlike traditional tools that re-discover information for every query. The system features an ingestion pipeline that extracts data, a knowledge store for persistent storage, and a retrieval layer that uses a search index and LLM reasoning to answer questions. AI

A developer has created an open-source social media management tool called BrightBean Studio, designed to offer features similar to paid services like Sendible and SocialPilot without cost or limitations. The platform allows users to plan, compose, schedule, and monitor content across numerous social media channels from a single dashboard. It can be deployed locally, via Docker, or on cloud platforms like Heroku and Render, with options for both PostgreSQL and SQLite databases. AI

Users of Anthropic's Claude Code Pro Max plan are experiencing rapid quota exhaustion, with some reporting their 5x quota being depleted in as little as 1.5 hours. The issue appears to stem from how "cache_read" tokens are accounted for against rate limits, potentially counting at full rate instead of the expected reduced rate. This behavior negates the benefits of prompt caching and causes quota to deplete quickly, even with moderate usage, exacerbated by background sessions also consuming shared quota. AI

Eve is a new AI-powered assistant designed to manage various work-related tasks, including research, writing, and coding. The tool offers over 100 built-in skills and aims to automate administrative duties. Eve is backed by Y Combinator and is preparing for a launch next week. AI

DemandSphere has migrated its website from WordPress to Jekyll, a static site generator, to improve speed and development flexibility. The company found WordPress limiting and difficult to find developers for, despite the general trend towards AI and markdown. Claude Code was instrumental in analyzing and filtering content, and assisting with the migration of blog posts and images, enabling the team to complete the project efficiently. AI

A Vercel plugin for Claude Code was found to be collecting extensive user data, including all typed prompts and bash commands, even for projects unrelated to Vercel. The plugin uses prompt injection to ask for telemetry consent, rather than a standard UI element, and misrepresents the scope of data collected as "anonymous usage data." While Vercel has since addressed the concerns and removed the telemetry code, the initial implementation raised significant privacy issues regarding data collection and user consent. AI

Relvy, a startup from the Y Combinator (YC) Winter 2024 batch, has launched its on-call runbook automation product. The platform aims to streamline incident response by providing automated runbooks. This launch targets engineering and operations teams. AI

CSS Studio is a new design tool that integrates AI agents to assist in the creation of web interfaces. The tool allows users to design visually using sliders and iterate on designs by generating multiple variants from a single prompt. It also enables one-click adaptation of layouts for different screen sizes, including mobile, with AI agents handling each task independently to maintain workflow efficiency. AI

Users are reporting an issue with Claude Code on Windows where the login process fails due to an OAuth timeout error. This problem prevents users from accessing the application, with the error message "OAuth error: timeout of 15000ms exceeded" appearing after attempting to log in via Google. The issue appears to be present in version 2.1.92 of Claude Code and affects users on Windows, particularly those using WSL. AI

Pace is a new application that connects users' wearable device data to Anthropic's Claude AI. This integration allows users to ask natural language questions about their health and fitness metrics, receiving personalized insights that go beyond raw numbers. The service aims to bridge the gap between collected data and actionable understanding, with plans for a companion app to visualize trends and comparisons. AI

Google's Gemma 4 model family, particularly the 26B-A4B variant, is now accessible for local inference on consumer hardware like MacBooks. This mixture-of-experts model activates only a fraction of its parameters per inference pass, enabling it to achieve quality comparable to much larger dense models while requiring significantly less memory and computational power. LM Studio's latest update, version 0.4.0, introduces a headless CLI, facilitating local setup and use of Gemma 4 and other models without a graphical interface. AI

A new project called Nanocode has been released, aiming to provide a high-performing Claude Code solution for $200. The project is built using JAX and is optimized for TPUs, suggesting a focus on efficient and powerful execution. AI

Cabinet is a new open-source, self-hosted AI-first operating system designed for managing personal and small business workflows. It stores all data as markdown files on disk, eliminating vendor lock-in and subscription fees. The system utilizes a team of AI agents to assist with tasks such as market research, content creation, sales outreach, and knowledge management, aiming to streamline operations for solo founders, creators, and small businesses. AI

A new toolkit has been released that leverages AI to assist users with travel hacking, specifically for finding award flights and planning trips. The toolkit integrates with AI coding environments like Claude Code and Codex, allowing users to query award availability across multiple mileage programs and compare costs. It offers both free and paid features, with the latter requiring API keys for services like award search, which are securely managed through a setup script. AI

A plugin called Superpowers significantly enhances the productivity of using Claude Code for software development. It introduces a structured workflow that guides the AI through brainstorming, exploring options with tradeoffs, sketching a plan, drafting a design document, and finally implementing the code. This approach allows for more manageable feedback loops and greater confidence in the generated code compared to the default Claude Code experience. AI

A suite of community-developed tools and techniques are emerging to enhance the capabilities of Anthropic's Claude Code, an AI coding assistant. These tools range from diagnostic utilities like 'Her' that analyze session logs to improve understanding of agent behavior, to wrappers like 'Claudraband' and bash script implementations that offer more control and persistence for power users. Additionally, developers are creating structured reasoning skills and prompt engineering hooks to guide Claude Code's output more effectively, enabling non-coders to manage complex AI workflows and improve the reliability and safety of AI-generated code. AI

A new open-source PostgreSQL extension, pg_textsearch, has been released, offering advanced BM25 relevance-ranked full-text search capabilities. This extension integrates seamlessly with PostgreSQL's existing text search configurations and supports features like expression indexes for JSONB fields and multi-column searches. It is designed for high performance and scalability, with optimizations for top-k queries and partitioned tables, and is compatible with PostgreSQL versions 17 and 18. AI

A developer accidentally created a fork bomb using Claude Code, which caused their computer to freeze overnight and incurred a $600 API charge. The incident occurred when a "SessionStart" hook, intended to spawn background instances of a tool called "CC", recursively created new instances. This led to extreme memory pressure and a system lock-up, which inadvertently prevented further API calls and thus limited the financial damage. AI

Users of Anthropic's AI coding assistant, Claude Code, are reporting that they are hitting usage limits much faster than anticipated, disrupting their workflows. Anthropic has acknowledged the issue and stated it is their top priority to resolve. Some users suspect bugs within the system are inflating token costs, with one claiming to have found issues that could increase expenses by 10-20x. This comes shortly after Anthropic introduced peak-hour throttling and concluded a promotion that doubled usage limits outside of peak times. AI

Ahmed Nagdy has launched "Learn Claude Code," an interactive platform designed to teach users about Claude Code through practical exercises rather than traditional reading. The platform features 12 modules with integrated terminal simulators and config builders, allowing users to practice commands and generate project configurations without any initial setup or API key requirements. It also includes quizzes to verify understanding and various tools like a playground and cheat sheet to deepen user knowledge. AI

A personal workflow has been detailed for improving tax filing in Canada by integrating Claude's Code CLI with Obsidian. This method involves organizing tax documents within Obsidian and using Claude to extract relevant information from PDFs and update financial plans. The process aims to provide a detailed overview of one's tax situation, supported by contextual data and Python-based calculations. AI

Anvil is a new desktop IDE designed to streamline parallel agent development, particularly for users working with Claude Code sessions. The application addresses common frustrations such as context switching, agent coordination, and tracking agent progress. By managing agent lifecycles and context, Anvil aims to improve developer productivity and push the boundaries of agent programming. AI

The .claude/ folder within Claude Code projects serves as a central control panel for customizing the AI's behavior. This folder contains instructions, custom commands, permission rules, and session memory, allowing users to tailor Claude to specific team needs. There are two such folders: one at the project level for team-wide configurations committed to version control, and another in the user's home directory for personal preferences and local state. AI

Anthropic has released "Routines" for Claude Code, a new feature allowing developers to automate coding tasks. These routines can be triggered by schedules, API calls, or GitHub events, enabling automated backlog maintenance, alert triage, code reviews, and deployment verification. A third-party tool, Claude Observe, has also emerged to provide real-time observability into these autonomous agent sessions, addressing the current opacity of multi-agent workflows. AI

Anthropic's new bug-hunting AI model, Mythos, has reportedly been accessed by unauthorized individuals through a third-party vendor environment, despite Anthropic's efforts to control its release. Early assessments suggest that while Mythos is efficient at finding vulnerabilities, its capabilities may not fully live up to the significant hype and concern generated by the company. The incident highlights the challenges of managing sensitive AI model releases and raises questions about the actual severity and exploitability of the vulnerabilities it has identified. AI

Anthropic has announced updates to its session limits for Claude, its AI assistant. The company is implementing new measures to manage usage and ensure a stable experience for all users. These changes are intended to prevent abuse and maintain the quality of service. AI

Anthropic has reportedly reduced the daily usage limits for its Claude Code assistant. The adjustment appears to have decreased the 5-hour limit for code-related tasks, though the exact nature and duration of this change are not fully detailed. This modification may impact developers relying on Claude Code for their coding workflows. AI

Orloj has released an open-source infrastructure-as-code platform for managing multi-agent AI systems. The tool allows developers to define agents, tools, models, memory, and other components using YAML and GitOps principles. Orloj aims to provide a declarative stack for building, operating, governing, and observing complex agentic systems, treating them like traditional software infrastructure. AI

A recent hackathon hosted by Anthropic showcased the power of domain experts in creating AI applications, with a lawyer winning first place for his AI-powered ADU permit assistant. The event highlighted how individuals with specialized knowledge can now build functional software without extensive coding experience. However, the article argues that while domain expertise is crucial for initial development, it doesn't guarantee a product's long-term viability, auditability, or robustness against edge cases and evolving regulations. AI

This is a Show HN post for a tool called Qretests, which aims to automate the creation of end-to-end test code by recording manual quality assurance flows. The tool integrates directly into a user's existing code repository, suggesting it's designed for developers and QA engineers looking to streamline their testing processes. AI

A building design consultancy owner has developed an AI agent, dubbed 'the talker,' to handle client inquiries and replace the need for junior staff. The agent, built over four months using a duct-taped stack including DeepSeek-R3, aims to improve responsiveness through techniques like 'Eager RAG' and by omitting persistent databases. The developer highlighted a recent interaction where the AI successfully defended its business model against a questioning architect, though the AI's aggressive tone has since been toned down. AI

A new tool called Stardrift.ai has been launched to help travelers predict which flights will offer Starlink internet service. The tool aggregates data on airlines, aircraft types, and specific aircraft tail numbers to estimate the likelihood of Starlink availability. This initiative is driven by the increasing affordability of rocket launches, making technologies like Starlink more accessible. AI

Kita, a startup founded by Carmel and Rhea, has launched a new product designed to automate credit review for lenders in emerging markets. The system utilizes Visual Language Models (VLMs) to process diverse and often unstandardized financial documents, a task that current OCR and document AI tools struggle with. Kita's platform extracts structured financial data, detects fraud, and verifies information through cross-document checks and historical data, aiming to improve the speed and accuracy of underwriting. AI

A new command-line tool called Claudetop offers real-time cost tracking for Anthropic's Claude models, addressing a lack of visibility into token usage and expenses. The tool provides detailed breakdowns of costs per session, model, and project, aiming to prevent unexpected billing surprises. Additionally, discussions highlight Claude's strengths in instruction following, coding, and long-form writing compared to competitors like GPT-4o, while also noting its larger context window and cleaner API for developers. AI

This resource provides 101 interactive exercises for learning the Pandas library for data analysis. Users can run and edit all code blocks directly within their web browser without any installation required. The exercises cover fundamental operations such as importing Pandas, creating Series from various data structures, manipulating indices, combining Series into DataFrames, and calculating descriptive statistics. AI

Spine Swarm, a Y Combinator-backed startup, has launched a platform that utilizes over 300 AI agents to conduct research and generate client-ready documents. The system claims to achieve the top ranking on Google DeepMind's DeepSearchQA benchmark, outperforming models like Claude and ChatGPT. Spine's approach involves parallel agent swarms that handle distinct workstreams, passing structured outputs to create deliverables such as reports, presentations, and spreadsheets. AI

IonRouter has launched a new inference service designed for high throughput and low cost, utilizing its proprietary IonAttention engine. This engine is capable of multiplexing multiple models on a single GPU, enabling rapid model switching and real-time traffic adaptation. The service supports various open-source models and fine-tunes, offering per-second billing and minimal cold start times, making it suitable for applications like robotics and real-time video analysis. AI

Axe is a new command-line interface tool designed to manage and execute AI agents, drawing inspiration from Unix philosophy for focused, composable functionality. It allows users to define agents with specific skills using TOML files, enabling them to be chained together or triggered by standard system tools like cron or git hooks. Axe supports multiple LLM providers, including Anthropic and OpenAI, and offers features such as persistent memory, sub-agent delegation, and structured JSON output for scripting. AI

Manuel Schipper has developed 'nah', a Python-based tool designed to provide deterministic safety guards for coding agents like Claude Code. Unlike prompt-based or allow/deny list approaches, nah maps tool calls to a taxonomy of intents, adding context such as project details and sensitive file locations. It then applies user-defined configurations and classifiers to either allow, block, or ask for clarification before an action is executed, aiming to prevent security risks like credential leaks or destructive commands. AI