news - News

Ai
blog_img

Docling v2.91.0 Improves Document Image Extraction and Parsing Reliability

Docling v2.91.0 adds DOCX VML image extraction and fixes METS-GBS validation, EasyOCR downloads, VLM templates, HTML image handling, backend behavior, and PPTX notes mapping.

Ai
blog_img

Daytona v0.169.0 Strengthens Sandbox Networking and Build Controls

Daytona v0.169.0 adds runtime sandbox network updates, per-runner declarative build sandbox caps, a Python SDK WebSocket concurrency limit, and key security hardening fixes.

Ai
blog_img

GitHub Spec Kit v0.8.0 Adds Template Composition and Skills-Based Scaffolding

GitHub Spec Kit v0.8.0 introduces composition strategies for templates and scripts, adds skills-based scaffolding support for Copilot integrations, and improves install and overwrite behavior.

Ai
blog_img

LangChain OpenAI 1.2.0 Fixes ChatOpenAI Streaming Hangs and Tightens Test Workflows

LangChain has released langchain-openai 1.2.0 with a key fix to prevent silent streaming hangs in ChatOpenAI, alongside CI and integration test workflow updates.

Ai
blog_img

Hermes Agent v0.11.0 Adds React TUI, Bedrock Support and New Transport Layer

Hermes Agent v0.11.0 introduces a React/Ink terminal interface, a new pluggable transport architecture, native AWS Bedrock support, expanded provider paths, and broader plugin and messaging platform support.

Ai
blog_img

Oh My OpenAgent v3.17.5 Tightens Runtime Compatibility and Release Stability

Oh My OpenAgent v3.17.5 introduces compatibility updates across package detection, plugin and config handling, delegate-task behavior, runtime registration, and safer install and publish workflows.

Ai
blog_img

Google Gemini CLI v0.39.1 Strengthens Workspace Trust and Shell Safety

Google Gemini CLI v0.39.1 adds stricter workspace trust enforcement in headless mode, secures .env loading, and improves shell command validation with a core tools allowlist.

Ai
blog_img

Sim v0.6.56 Adds AgentPhone Integration and Better Subagent Visibility

Sim v0.6.56 introduces AgentPhone integration, bounded subagent log views, improved file sorting, and fixes across API responses, table resizing, retention controls, and image handling.

Ai
blog_img

Open WebUI v0.9.2 Expands Document Extraction, Firecrawl Reliability and Admin Controls

Open WebUI v0.9.2 adds PaddleOCR-vl document extraction, direct Firecrawl v2 support, customizable calendar reminders, flexible API key headers, OAuth session disconnection, and clearer source visibility in chat.

Ai
blog_img

OpenClaw v2026.4.23 Expands Image Generation and Sub-Agent Context Controls

OpenClaw v2026.4.23 adds Codex OAuth and OpenRouter image generation support, forked context for sessions_spawn, configurable generation timeouts, and tunable local memory embedding context size.