AI.info
AI Tools Updates — Page 15
What changed in the AI tools this week — dated, from the makers’ own changelogs.
Em inglês
- Keploy — v3.6.229
- OpenHands — v1.38.0
Added editable plugin configuration in automation help, conversation budget meters, and automation presets for agent profiles.
- Palantir AIP — Test media upload workflows on global branches
Media upload actions can now run on a global branch, writing uploads to the backing media set on your branch.
- Palantir AIP — Explore complex connections across your Ontology with Search Around V2
Search Around V2 is generally available in Vertex for building searches across multiple object and link types, following repeating patterns, and finding connections…
- Palantir AIP — Preview SuperRepo applications with live Ontology data using remote preview
REMOTE_ONLY mode runs OSDK React applications and functions locally against live objects and links in the Ontology, read-only and using your own permissions.
- Augment Code — Auggie CLI 0.37.1 Release Notes
0.x installs with auto-update enabled now upgrade to Auggie 1.x.
- Augment Code — Auggie CLI 1.2.0 Release Notes
Auggie 1.2.0 rewrites the Auggie CLI, built on the pi coding agent.
- Lunit INSIGHT CXR — Lunit to Supply AI-Powered Imaging Solutions to Saudi Arabia's Ministry of National Guard Health Affairs
Lunit to supply AI-Powered Imaging Solutions to Saudi Arabia's Ministry of National Guard Health Affairs
- n8n — n8n@3.2.0
Fixed fetch URL handling, workflow history, package compatibility, gzip filenames, redaction, webhook decoding, and Gemini model pickers.
- OpenRouter — Introducing the Extractions API
Launched the Extractions API for pulling typed fields out of documents and images.
- Hugging Face — Decision Models in Hugging Face
Decision models are available in Hugging Face.
- Hugging Face — Pay for Hugging Face Hub with your Google Cloud Account
Hugging Face Hub can be paid for with a Google Cloud account.
- Figma AI — New Figma Motion updates: physics based springs, style and Lottie export previews
Figma Motion adds animation style previews, Lottie export previews, and physics-based spring controls.
- GitHub Copilot — C++ language server now auto-configures MSBuild builds
- GitHub Copilot — Gemini 4 Carbon in GitHub Copilot
- GitHub Copilot — GitHub App support for Enterprise Copilot seat management
- ChatGPT — Business Hub in ChatGPT
Claim a business, publish updates, connect a product feed, view insights, and create a business agent in Business Hub.
- Ollama — v0.40.17
Improved ChatGPT CLI restore and restart, OpenCode settings, Vulkan mmap, concurrent MLX generation, and updated llama.cpp and MLX.
- Riverside.fm — Find your best mic
Let Riverside pick the mic that sounds best before you join the studio.
- ThoughtSpot Analytics — You can now change a parameter’s value just by asking Spotter. The answer updates immediately.
- Biela.dev — v5.0.0: Private previews, What's new and deeper research
Your app previews are now private by default and easy to share on your terms: by sign-in, email code, PIN, share link or a temporary Public switch.
- ElevenLabs — Scribe v3
Scribe v3 and Scribe v3 Live are now available in the Speech to Text API for recorded audio and real-time transcription.
- ElevenLabs — Workflows as code
ElevenAgents workflows can now be built in code with the Workflow builder in the Python and JavaScript SDKs.
- ElevenLabs — Unity SDK
The ElevenAgents Unity SDK is now available as the io.elevenlabs.agents package.
- Amazon Bedrock — Added Claude Fable 5.2 and Claude Mythos 5.2
Added model cards for Claude Fable 5.2 and Claude Mythos 5.2; both are available through US and Global cross-Region inference.
- Hasura DDN — NDC Spec
Define canonical capability-gated relational streaming errors.
- Hasura DDN — NDC Clickhouse
Updated webpki-roots, rustls-webpki, quinn-proto, rustls, tokio, anyhow and rand; require the eMudhra distrust CA bundle in the runtime image.
- Hasura DDN — NDC MongoDb
Share immutable MongoDB configuration across router-state and per-query clones instead of copying the full configuration, reducing memory usage.
- Leonardo AI — GPT Image 2.6 is now on Leonardo.Ai
Refine one image, edit by edit, with GPT Image 2.6 Flare and Sunburst and up to 16 image references.
- Higgsfield AI — Seedream 5.5
Seedream 5.5 is now on Higgsfield and can generate images with transparent backgrounds.
- Higgsfield AI — Fable 5.2
Fable 5.2 is now available in Supercomputer on Higgsfield.
- Jellypod — Math Video Style
Make math videos and shorts with step-by-step equations and graphs that draw along with the narration.
- Taskade — v6.308.0: Public Agent Chat Privacy & Cleaner Kit Copies
Public agent visitors no longer see tool inputs and outputs or reasoning steps; workspace members still see every step.
- Wordly — Portal 4.18.0
Added "Download Presenter QR Code" and "Attend now" buttons to the "Ways to Join" page; fully translated the Transaction report export.
- Wordly — Try It Now 2.1.1
We've made improvements in accessibilty and fixes for increased security.
- Wordly — Display 2.6.1
Fixed an issue where a display could show the same content two or more times; made improvements in accessibilty; made fixes for increased security.
- Gensyn Testnet — open-8b: an auditable model, in training
Gensyn is training open-8b to be an auditable model.
- Slack AI — Security Guidance
This release includes security improvements.
- Auphonic — Sync Tracks: Automatic Alignment for Your Multitrack Recordings
- AirOps — Fact Check and ChatGPT Shopping Insights
Check whether AI answers describe your brand accurately, see products ChatGPT surfaces, and reuse content processes in Campaigns.