ai-agents

Kitten TTS ships sub-25MB expressive TTS models, AWS unveils V-RAG for grounded video generation, and Thomson Reuters explains why agentic workflows win in 2026

Mastra 1.14.0 lands AI Gateway tool support and more reliable observational memory, plus AWS ships Strands Evals for agent evaluation and an AI A/B engine, as 


THOR AI accelerates materials physics simulations, NanoClaw partners with Docker for secure AI agents, and Ai2 ships open-source sim-only robotics tools, as of

GitAgent proposes a git-native agent spec, Hyperspace AGI launches a distributed peer-to-peer AGI network, and ARIS automates Claude Code research workflows. As

Spine Swarm’s collaborative canvas agents, Captain’s automated RAG for files, and Context Gateway’s context compression show how production AI agents are matur​

ESP32-scale agents, sovereign AI governance, and Wayfair’s production OpenAI stack show where serious agentic engineering is headed next.

Zendesk buys Forethought for self-learning AI agents as LangChain ships autonomous context compression and OpenAI expands agent runtimes. What this means for AI

Hugging Face pushes million-token training with Ulysses SP, AWS ships serverless Nemotron 3 Nano, and LangGraph 1.1 adds type-safe streaming for serious AIagent

Cut MCP token waste with mcp2cli, boost tool-calling with Mastra 1.10.0, and explore open-source high‑fidelity speech with MOSS‑TTS for production AI agents.

AI agents now run a one‑person company for $0, coordinate in 13‑agent swarms, and power cyberattacks. Plus new orchestrators, semantic merge, and OpenClaw play‑