HTTP
PulseAugur coverage of HTTP — every cluster mentioning HTTP across labs, papers, and developer communities, ranked by signal.
11 day(s) with sentiment data
-
Portable AI agents can now run from a 340MB USB stick package
A new project called norax-portable enables the creation of self-contained AI agents that can run on any x86_64 Linux machine from a USB stick. The package, which is only 340MB, includes Python, Ollama (CPU-only), and m…
-
AWS and Cisco introduce agentic overlays for legacy service integration
AWS and Cisco have collaborated on a new approach called "agentic overlays" to integrate legacy enterprise services with emerging Agent-to-Agent (A2A) communication standards. These overlays act as thin wrappers, transf…
-
Modal launches ultra-low-latency servers for high-performance applications
Modal has introduced a new feature called Modal Servers, designed to provide ultra-low-latency server hosting for applications requiring high performance, such as LLM inference for interactive agents. This new offering …
-
OpenAI launches AI security initiative amid concerns over LLM agent flaws and data privacy
OpenAI has launched DayBreak, an initiative focused on enhancing AI security and protecting models from cyber threats. Concurrently, researchers have identified a critical flaw in LLM agents called 'constraint decay,' w…
-
Cloudflare partners with major browsers on bot detection
Cloudflare is collaborating with major browser developers, including Google Chrome, Firefox, Safari, and Microsoft Edge, to implement a new system for distinguishing between human users and automated bots. This initiati…
-
AI agents fail silently when tool schemas change without error
A subtle but critical failure mode in AI agent tool usage has been identified, where tool schemas can change without triggering errors. Agents may continue to send requests using outdated schemas, leading to silently in…
-
AI agents vulnerable to indirect prompt injection via web scraping
An AI agent can be tricked into executing malicious commands by scraping web pages that contain hidden instructions. This indirect prompt injection occurs when an agent processes data from external sources, such as a we…
-
Cursor IDE's HTTP Streamable server fails for remote connections
Users of the Cursor IDE are reporting issues with its HTTP Streamable (oAuth) server, which has stopped working for remote MCP connections. This functionality, which previously worked, now fails specifically within Curs…
-
Smeldr aligns with Anthropic's MCP server best practices
Smeldr's author found that Anthropic's recommendations for building production-ready MCP servers align with their own design decisions. The author highlights how Smeldr's implementation of remote HTTP servers, intent-gr…
-
Cloudflare introduces 402 Payment Required for web crawlers
Cloudflare has begun implementing the HTTP 402 Payment Required status code for its Pay-Per-Crawl service, a move that significantly alters how web crawlers interact with websites. Previously, crawlers faced 200 OK for …
-
42 LLMs benchmarked for speed: Smaller models often faster
An independent tracker named ollamatps.com has benchmarked 42 large language models (LLMs) to measure their actual response speed, distinguishing between Time to First Token (TTFT) and Tokens Per Second (TPS). The bench…
-
redb.Route integrates LLMs as endpoints, unifying AI with existing frameworks
The redb.Route integration framework has released version 3.1.0, introducing two new transports: redb.Route.Llm and redb.Route.Exec. The LLM transport allows developers to treat language models as addressable endpoints,…
-
SSH tunnels offer simpler encryption for local AI model traffic
Running local AI models across multiple machines often exposes data over unencrypted HTTP, posing a security risk on shared networks. The author proposes using SSH tunnels as a simpler alternative to setting up full TLS…
-
Pilot Protocol offers agents a dedicated session layer, improving MCP tool use
A new protocol called Pilot is emerging to address limitations in the current agent communication stack, particularly for tools like MCP. While MCP excels at the application layer for exposing tools to LLMs, it relies o…
-
AI agents waste context on HTML; HTTP spec offers a fix
AI agents waste a significant portion of their context window processing HTML scaffolding instead of actual content. A solution has existed within the HTTP specification for 27 years, but it has been largely unutilized.
-
LLMs enable taxonomy-agnostic PII annotation in HTTP traffic
Researchers have developed a novel pipeline utilizing Large Language Models (LLMs) to automatically identify and annotate Personally Identifiable Information (PII) within HTTP traffic. This method aims to overcome the l…
-
x402 protocol enables pay-per-call MCP servers with USDC
The x402 protocol enables pay-per-call functionality for AI model servers, addressing the issue of unlimited free access in existing setups. It revives the HTTP 402 Payment Required status code to facilitate instant USD…
-
AI agents gain cross-platform communication with new A2A Protocol v1.0
Microsoft has released version 1.0 of its A2A Protocol, a stable and production-ready standard for cross-platform communication between AI agents. This protocol aims to eliminate the need for custom integration code by …