AI.info
AI Tools Updates — Page 29
What changed in the AI tools this week — dated, from the makers’ own changelogs.
- Pinecone — Restoring a backup to a different region is generally available
Restore a serverless backup into a different region on the same cloud provider in the Pinecone console and API.
- Snowflake Cortex Analyst — Level of detail (LOD) expressions in semantic views (Preview)
Level of detail (LOD) expressions in semantic views
- ChatGPT — Outbound SIP calling in the Live API
Outbound SIP calling is now available to all organizations through the Live API. Connect GPT‑Live to a phone number through your SIP provider.
- Harness AI (AIDA) — Unified API Inventory: Accelerate Patching with Cross-Source
Unified API Inventory uses cross-source API correlation to connect runtime findings to code, documentation, and fixes.
- Harness AI (AIDA) — Harness October 2026 Updates: 150+ Features Shipped.
Harness shipped 150 features in October 2026, including the Cosmos Software Factory Agent, MCP sign-in with an identity provider, and FedRAMP 20x Class A.
- Ghostwriter — Redesigning Agent Studio around Ghostwriter
Agent Studio was redesigned around Ghostwriter, bringing conversation analysis, journey updates, and testing into one conversation.
- ON1 Photo RAW — ON1 Photo RAW 2027 Is Available Now: New RAW Engine & More
ON1 Photo RAW 2027 launches for Mac and Windows with a new RAW engine, improved NoNoise AI, enhanced masking, and more.
- Qodo — Qodo IDE plugin deprecation
The Qodo IDE plugin is no longer supported as of October 30, 2026. Use the Qodo Agentic Toolbox with a supported coding agent to review code locally.
- Qodo — Configure rule import sources and instructions
Specify file names and folders as rule sources, and provide instructions to guide which rules Qodo extracts and where they apply.
- Qodo — More control over review verbosity and publishing locations
The routing matrix shows where Qodo publishes each group. Saving a display mode or single row saves every row; Custom is now called Mixed.
- Qodo — Add context from other pull requests in the same repository
When you link related pull requests in the description, Qodo uses them as additional context for the review.
- Qodo — Repository access filtering for GitHub Enterprise Server
Enterprises can enforce repository-level access control in Qodo using GitHub Enterprise Server or GitLab permissions.
- Letta — Release v0.34.39
Allows agent messages to address a conversation by ID alone and attaches recovery consumers to live duplicates.
- Letta — Release v0.34.38
Streams subagent progress into task output files and logs every subagent task as JSONL, including remote workers.
- Letta — Release v0.34.37
Shows the configured monitor cap and reads it in seconds; resolves cron computer names or IDs to device IDs.
- Tines Agents — 🔐 SSH connectors now support password authentication
SSH connectors now support password authentication.
- Tines Agents — 🌐 Set custom domains for workflow webpage URLs. Configure in Settings > Domains
Set custom domains for workflow webpage URLs in Settings > Domains.
- Tines Agents — ✨ Auto-generate pull request descriptions for Git-synced workflows
Auto-generate pull request descriptions for Git-synced workflows.
- Tines Agents — 🔌 Filter connectors by type
Filter connectors by type.
- Tines Agents — 🔗 New Google Drive and Gmail OAuth connectors
New Google Drive and Gmail OAuth connectors.
- Higgsfield AI — Higgsfield Mangaka
Higgsfield Mangaka lets users create, edit, and export manga, comics, and webtoons from an idea or script.
- Meshy — Added ai_model: "meshy-n1" to the Retexture API.
Meshy N1 makes a 2K base color texture from one to four reference images; a task costs 10 credits.
- Jellypod — Direct YouTube Publishing
Publish videos and Shorts directly from the Jellypod editor to your YouTube channel.
- PromptLayer — Vibe Chat Daily Limits
Vibe Chat is available on paid plans, with a daily message limit per user. Trial workspaces get a smaller daily allowance.
- Taskade — v6.301.1: Stripe Restricted Keys & App Video Hotfix
Stripe connections accept restricted API keys, and videos in Genesis app previews can play full screen.
- Taskade — v6.300.3: Automation Branch & Webhook Security Hotfix
- Taskade — v6.301.0: Agent Team Roles & Keep Automations Going
Agent team members can have roles, and automation steps can stop or continue when a step fails.
- Modal — New Image Builder version 2026.10 with updated trust stores
Image Builder version 2026.10 includes refreshed trust stores following the recent eMudhra CA compromise and improvements to build caching.
- Sierra Insights — Redesigning Agent Studio around Ghostwriter
Agent Studio was redesigned around Ghostwriter to examine conversations, update journeys, and test fixes in one conversation.
- Sakana Fugu — 【開催レポート】Sakana AIシンポジウム:Jürgen Schmidhuberが語るAIのこれまでとこれから、日本から挑む言語モデル・世界モデル・Physical AI
自社で事前学習する言語モデルシリーズ「Sakana Sawara(仮称)」を発表し、最初のモデル「Sawara Small(仮称)」を開発する。
- Subframe — Edit selected elements with agent chat
Select an element on the canvas and ask agent chat to change it, and the agent edits exactly what you selected.
- Wordly — API 1.14.0
We've made vulnerability fixes for increased security.
- Honeycomb — Reply to selected text in Canvas
You can select text in Canvas and ask the agent about it in chat with the text attached as context.
- Honeycomb — More Canvas connectors: GitLab, LaunchDarkly, Stripe, CircleCI, Intercom, and Amp
The Canvas Agent can now connect to GitLab, LaunchDarkly, Stripe, CircleCI, Intercom, and Amp.
- Digits AI Invoicing — Add Internal Client IDs
Accounting firms can add optional Client IDs to client records, shown next to client names on the Firm Dashboard.
- Google AI Studio — Gemini 2.5 Flash deprecation
gemini-2.5-flash is deprecated and no longer available to select in Google AI Studio; starting December 1, 2026, requests route to gemini-3.5-flash-lite.
- Google AI Studio — Gemini 2.5 Pro deprecation
gemini-2.5-pro is deprecated and no longer available to select in Google AI Studio; starting December 1, 2026, requests route to gemini-3.8-flash.
- Omni — October 26, 2026
MCP tokens are enabled by default for eligible organizations; admins can add global AI context files and prefer apps; AI routines can deliver to webhooks.
- Recall — Recall release notes, 2 November 2026: Chat and MCP get smarter at finding your content, plus better support for Gmail and YouTube
Chat and MCP can find exact phrases and read whole cards; connected apps can be managed; Gmail saves and age-restricted YouTube support improved.
- World Anvil — World Anvil News: November 2026
World Anvil updated user profiles and introduced a new way to create worlds.