AI.info
AI Tools Updates — Page 43
What changed in the AI tools this week — dated, from the makers’ own changelogs.
Auf Englisch
- Snowflake Cortex Analyst — Carry over comments, tags, and privileges with COPY = ( … )
Carry over comments, tags, and privileges with COPY
- OpenHands — v1.34.0
Added per-conversation Docker runtimes and a Runtime settings page; automations can be filtered and edited by tag.
- Genspark — Genspark Media Studio: One AI Studio for Any Video
One AI Studio for Any Video
- JetBrains AI Assistant — The Terminal in JetBrains IDEs Now Runs on libghostty
Starting with 2026.3, the built-in terminal in JetBrains IDEs runs on libghostty.
- LangChain — 2026-10-19 - LangSmith Cloud update
A new sidebar is available in every workspace, and example creation now uses reference_outputs across the API and SDKs.
- Lovable — Redesigned sidebar
The sidebar organizes your projects and chats into Pinned, Projects, and Recent chats.
- Lovable — Connect your app to Scrive
Connect your Lovable app to Scrive for e-signatures and electronic identification (eID).
- Lovable — Events & hooks for your workspace
Workspace admins and owners on Business and Enterprise plans can extend their workspace with their own services using Events & Hooks.
- Lovable — Knowledge is now Instructions
Workspace knowledge and project knowledge are now called workspace instructions and project instructions. They work the same way.
- OpenRouter — CSAM Screening Across OpenRouter
OpenRouter will screen all inference requests for potential CSAM during processing.
- Hugging Face — Build specialized multimodal models with LFM3
LFM3 enables building specialized multimodal models.
- Runway — Runway Ads
Reads your best performers on Meta, Google and TikTok, generates variants for every placement and publishes the ones you approve.
- Figma AI — View and restore previous versions of your generative plugins and shaders
Editing generative plugins and shaders creates a new version, and previous versions can be revisited and restored.
- ChatGPT — iCloud Mail and Calendar plugins in ChatGPT
iCloud users can now connect Mail and Calendar to ChatGPT to find emails, catch up on your inbox, check your schedule, and update calendar events.
- Ollama — v0.40.12
Ollama adds Liquid AI's LFM3 model family and d1 decision models in a 2B size.
- mabl — 2.74.4 (2026-10-26)
Updated embedded CLI to 2.138.37; Fixed CSS assertions on a Find elements step leaving the expected value blank; Removed eMudhra CA support.
- PostHog — How Replay Vision helps users find (and solve) product problems
Replay Vision scans Session Replay recordings and summarizes findings based on parameters users set.
- Hightouch AI Decisioning — [New] Customer Studio
The Customer Studio agent opens beside the journey canvas, understands unsaved edits, and can apply changes for review.
- Gamma — Exports that stay closer to your design
PowerPoint exports preserve custom fonts in multilingual text, equations, and table emoji and layouts; tall charts fit PowerPoint and Google Slides more reliably.
- Gamma — Clearer credit balances
Check your balance from your home screen and Billing settings, with credit information and an out-of-credits state.
- Gamma — Introducing Gamma 5
The agent brainstorms, researches, and designs a presentation from a sentence, doc, PDF, image, or old deck.
- Gamma — Previews that match your format
New and updated Gammas show previews that reflect their format, including portrait documents, social posts, and 4:3 slides.
- Amazon Bedrock — Clarified when Advanced Prompt Optimization prompt shortening returns the original prompt
Clarified that if the original prompt scores at the maximum, or samples do not distinguish candidate prompts, the original prompt is returned.
- Amazon Bedrock — Clarified batch inference completion behavior and quotas
Clarified that on-demand requests are prioritized over batch inference jobs and job completion is not guaranteed; added job timeout values.
- Meshy — New Deprecations page: the lifecycle of every 3D generation model (active, legacy, deprecated, or retired) with its retirement date and recommended replacement, the history of every API deprecation, a
Lists the lifecycle of every 3D generation model, API deprecations, and deprecated parameters still in service.
- iSpring LMS — Create more natural, expressive voiceovers with Eleven v4
Eleven v4 is now available for AI voiceovers in scrollable courses, takes context into account, and supports more languages.
- Hightouch — [New] Customer Studio
The Customer Studio agent works in a side panel beside the journey canvas and can explain, analyze, and apply changes to journeys.
- LTX Studio — LTX’s Latest VFX Tools: Built for the Production Pipeline
LTX’s latest VFX launch brings new tools for the work that happens between an idea and a finished shot.
- Atlan — Context Agents now link glossary terms to the right tables, views, and columns
The Asset ↔ Term Linker agent proposes links between glossary terms and the right tables, views, and columns.
- Guru — Set chat thread retention for each Knowledge Agent
Knowledge Agent owners can set chat thread retention per agent, from 1 to 90 days.
- Guru — Mermaid diagrams in cards and Chat
Guru now renders Mermaid diagrams in cards and Chat, including diagrams in Knowledge Agent responses and content added through MCP.
- Guru — DLP masking now covers what users type
Workspace-level DLP rules now mask user input in search, chat, MCP requests, and questions asked from Slack and Teams.
- Cline — sdk/ui/v0.2.0
Released @cline/ui version 0.2.0.
- Cline — sdk/shared/v0.2.0
Released @cline/shared version 0.2.0.
- Cline — sdk/agents/v0.2.0
Released @cline/agents version 0.2.0.
- Cline — sdk/llms/v0.2.0
Released @cline/llms version 0.2.0.
- Cline — sdk/core/v0.2.0
Released @cline/core version 0.2.0.
- Cline — SDK v0.2.0
SDK v0.2.0 reattaches to live sessions, tracks connected capability clients, and retries turns when the connection drops after reasoning.
- Capacities — Whiteboards, global graph view, and polish on the basics
Added whiteboards and a global graph view, with workflow and usability improvements.
- vLLM — v0.33.0rc0
Skip adaptive verification cost profiling in CUDA.