AI.info
AI Tools Updates — Page 13
What changed in the AI tools this week — dated, from the makers’ own changelogs.
En inglés
- Harvey AI — Exports Page for Admins
A new Exports page in Command Center lists the Workspace History, Analytics, and Workspace Vaults exports you request and lets you download each finished file until it…
- Smartlead — Faster Mailbox Deletion & Easier Mailbox Management
Mailbox deletions are permanent after 1 hour, with a Recycle Bin to restore them during that window.
- Higgsfield AI — Gemini 4 Carbon
Gemini 4 Carbon is now available in Supercomputer on Higgsfield.
- Meshy — Added skeleton_template to Rigging .
It delivers a biped character on the skeleton your engine or tool expects; Animation tasks use the rig's skeleton.
- Meshy — Added include_fingers to Rigging .
Set it to true to add finger bones to a biped character's hands, on any skeleton.
- Meshy — meshy-7 retires on December 31, 2026 as the ai_model of Text to 3D preview tasks, Image to 3D , and Multi-Image to 3D .
From that date, requests that send ai_model: "meshy-7" to these endpoints return 400; switch to meshy-7.1.
- Culture Amp AI Coach — AI Follow-Ups Now Included in Sensitive Comment Management
Sensitive Comment Management now scans and flags answers from AI Follow-Up conversations alongside other comments.
- Picsart AI Photo Editor — Gemini Omni 1.2 Flash vs 1.1 Flash: longer clips, closer references
Gemini Omni 1.2 Flash is available in Picsart alongside Gemini Omni 1.1 Flash.
- Picsart AI Photo Editor — Seedream 5.5 is now in Picsart: realistic images that hold up edit after edit
Seedream 5.5 is now available in Picsart for making and editing images.
- Hyperscience Platform — Upgrades with flows that use unsupported blocks
Upgrades no longer get stuck at the init step; unsupported flows and flows that call them as subflows are undeployed.
- Taskade — v6.309.0: One-Click App Connections & Date Fields Over MCP
Connect apps from the Connections page, and set task date fields through MCP and the API.
- Subframe — Fixes & improvements
Undo and redo after moving many pages undo the right edit and stay fast. Designs that fill the browser window fill their whole page on the canvas.
- Subframe — Subframe plugin for ChatGPT and Codex
Install the Subframe plugin from ChatGPT's plugin directory to design with Subframe in ChatGPT and build your designs in Codex.
- Deepgram — Deepgram Self-Hosted November 2026 Release (261110)
Container images include the self-hosted API, engine, license proxy, and billing.
- Culture Amp Engage — AI Follow-Ups Now Included in Sensitive Comment Management
Sensitive Comment Management now scans and flags answers from AI Follow-Up conversations alongside other comments.
- v0 by Vercel — Use your ChatGPT Business plan in v0
ChatGPT Business plans now work in v0, alongside Plus and Pro; eligible AI responses count towards your ChatGPT plan instead of v0 credits.
- v0 by Vercel — v0 Security Check
Ask v0 to review your changes for security issues like missing auth checks, injection, and SSRF.
- Factory Droids — CLI v1.7.1, Desktop v1.7.1: Stale thinking no longer fails the next message in a session
Droid continues a session after earlier thinking goes stale instead of failing the next message.
- Factory Droids — CLI v1.7.0, Desktop v1.7.0: Session sidebar performance, spec and merge commit fixes, and clearer sign-in and token messages
Streaming sessions no longer redraw the whole sidebar, and Cloud Machine notices point to Environments as the replacement.
- Honeycomb — AI Ecosystem: now GA
AI Ecosystem is now generally available, with fleet performance and cost metrics plus Agent Timeline insights.
- Digits AI Invoicing — Introducing Digits Pro 🚀
Digits Pro is now available with AI Bookkeeping, Agentic Close, Budgets, Custom Reports, and Dashboards.
- Digits AI Invoicing — New General Ledger Export Format
Financial package exports now include a General Ledger Excel file with transactions grouped by category and ending balances.
- Digits AI Invoicing — Budgets Are Here
Digits Pro now supports budgets, including Excel budget uploads and Budget vs. Actuals tracking.
- Google AI Studio — Gemini 4 Carbon ( gemini-4-carbon ) generally available (GA)
Released Gemini 4 Carbon (gemini-4-carbon), built for agentic coding, multimodal knowledge work, scientific reasoning, and cybersecurity defense.
- Flowith Canvas — Seedream 5.5 is in Early Access on flowith NOW
Available in Early Access for image generation and editing; edit people, products, or scenes across multiple rounds of iteration.
- Flowith Canvas — Claude Fable 5.2 joins flowith
Now live on flowith; handles reports, sources, data, proposals, charts, screenshots, and other visual materials alongside text.
- Mixedbread Rerank — Cheaper Ingestion and Storage
Ingestion and storage rates decreased, with the new rates applying automatically from the current billing period.
- Decagon Voice — How we taught our agents to recognize other agents
Taught agents to recognize other agents.
- Agent Neo — Seedream 5.5 is in Early Access on flowith NOW
Seedream 5.5 is now available in Early Access on flowith for image generation and editing.
- Agent Neo — Claude Fable 5.2 joins flowith
Claude Fable 5.2 is now live on flowith, with reasoning and visual understanding for longer, more intricate tasks.
- Speakeasy SDK Generation — Dashboard sign-ins stay bound to the browser that started them
- Speakeasy SDK Generation — Agents get a dedicated MCP sign-in tab, and their Permissions page leads with servers
Agents get a dedicated MCP sign-in tab, and their Permissions page lists reachable servers first.
- Weaviate — v1.40.5 - Alter schema reindex property improvements, Weaviate Embeddings module Fix
- Weaviate — v1.39.16 - Alter schema reindex property improvements, Weaviate Embeddings module Fix
- Weaviate — v1.38.25 - RBAC roles & Weaviate Embeddings module Fixes
Fixed RBAC roles and the Weaviate Embeddings module.
- Fern Docs — Remote paths
You can now serve a path prefix on your Fern docs domain from another HTTPS service you run.
- Gemini Enterprise Agent Platform — Cross-project governance for semantic governance policies
Use semantic governance policies in a multi-project architecture to centrally govern agents deployed across multiple Google Cloud projects.
- Gemini Enterprise Agent Platform — Provisioned Throughput supports more detailed order statuses
Provisioned Throughput supports more granular order statuses in the Google Cloud console in General Availability (GA).
- Ask Sage — Ask Sage v4.1.0
Adds an option to require multi-factor authentication, automatically retires unused API keys, and adds new models.
- Decagon AI Agent Platform — How we taught our agents to recognize other agents
Taught our agents to recognize other agents.