AI.info
AI Tools Updates — Page 107
What changed in the AI tools this week — dated, from the makers’ own changelogs.
Auf Englisch
- AirOps — New OpenAI and Anthropic models plus Ads insights
Compare YouTube citations across AI platforms, browse ChatGPT ads, and create with GPT-6 Sol, GPT-6 Luna, and Claude Opus 5.5.
- Gemini Enterprise Agent Platform — Anthropic's Claude Sonnet 5.5
Claude Sonnet 5.5 is available in Model Garden.
- Gemini Enterprise Agent Platform — Provisioned Throughput: Support for changing order scope and increasing term
Provisioned Throughput supports changing an order's scope or increasing its term through a supersede and replace operation.
- Gemini Enterprise Agent Platform — Gemini 3 models supported by the Interactions API in Preview
Gemini 3 models are supported by the Interactions API in Preview on Gemini Enterprise Agent Platform.
- Ask Sage — v3.9.4
Fix — service startup / database-driver regression
- Ask Sage — v3.9.5
Security maintenance (dependency & base-image updates); vector-database upgrade
- Distro — 🕸️ Web Experiences: new nodes, double ownership, and a narrower Announcement
Web Experiences adds routing nodes and double ownership support in more flows, plus CSS styling for the Announcement bar.
- Devin — September 25, 2026
Session status groups now separate Working, Ready, Blocked and Inactive; view preferences sync across browsers and devices.
- Agent Studio — Ghostwriter: When AI goes from tool to teammate
Ghostwriter now works proactively and asynchronously alongside teams in Slack and Teams.
- Clay — Product Roundup Week of September 21, 2026
Sync company lists to LinkedIn; control which IP addresses members can reach Clay from; trace connection usage; create workspaces in about a second.
- UiPath Maestro — Standardized layout for the execution panel's Inputs, Outputs, and Variables sections
Inputs, Outputs, and Variables sections now share a unified layout, with a title and search field for each.
- UiPath Maestro — Flow validation errors now appear in the VS Code Problems panel
Validation errors and warnings now appear in VS Code's Problems panel alongside your project's other diagnostics.
- UiPath Maestro — Create Maestro Flow projects in Studio Desktop
A Studio Desktop solution can now include Flow projects alongside your automation projects, created without leaving Studio.
- Concierge — 🕸️ Web Experiences: new nodes, double ownership, and a narrower Announcement
Added Assign To and Update Ownership nodes, new Scheduling Flow node placements, double ownership routing, and Announcement bar CSS styling.
- Claude Code — 2.1.284
Added Claude Sonnet 5.5 as the default Sonnet model on the Anthropic API, with 1M context and new pricing.
- Cresta Knowledge Agent — Handoff summary support for cold transfers
A concise handoff summary is generated when an agent transfers a call without context and is available in Closed Conversation view in Director.
- Cresta Knowledge Agent — Mocking API
Register an external API endpoint, review captured invocations, and promote an invocation into a mock rule for test cases.
- Cresta Knowledge Agent — Batch post-call translation
Automatically translate post-call conversations from 30+ languages into English in bulk.
- Cresta Knowledge Agent — Advanced Topic Discovery
Topic Discovery V3 updates LLM-generated clusters to produce conversation reasons at every level, including drill-down views.
- Cresta Knowledge Agent — AI feedback popover includes examples
The feedback popover shows examples of good and bad feedback to guide what to write.
- Cresta Knowledge Agent — New environments and deployments tabs
Manage agent configurations across testing, staging, and production with environment variables, secrets, revisions, traffic splits, and runtime settings.
- Cresta Knowledge Agent — Create test cases directly from AI Feedback
Users can create a test case directly from a feedback item; the flow generates a test case and evaluator based on the conversation and feedback.
- Cresta Knowledge Agent — Conductor (for developers)
Build, test, and optimize AI agents, including prompts, tools, and orchestration, from a local no-code interface guided by Cresta best practices and conversation data.
- Devin Desktop — September 25, 2026
Sidebar groups sessions by status, preferences sync across devices, and keyboard shortcuts navigate sessions and submit dialogs.
- Intercom Fin — Choose bigger Tooltips that are easier to see and click
Set each Tooltip to Small, Medium, or Large; the size applies to icons, animated beacons, and labels.
- Intercom Fin — Merge duplicate users into one profile
Merge one user into another from the conversation sidebar, control who can merge with a teammate permission, or merge users through the REST API (Preview).
- Fireworks AI — New training model: DeepSeek V4.1 Flash
DeepSeek V4.1 Flash is now available for LoRA training on the Dedicated Training API, with up to 262K context.
- OpenHands Cloud — openhands: v0.72.0
Added managed LiteLLM providers for OpenAI, DeepSeek, Mistral, Groq, OpenRouter, and Gemini, and made MinIO storage configurable for new installs.
- Persado Optimize — [PRODUCT UPDATE] Three new ways to use Persado
Three new entry points this quarter: Persado MCP inside Claude and Copilot, Persado Create from brief to production-ready, and a free Market Scan
- Pinecone — Released Go SDK v7.0.0
Uses API 2026-07 and adds support for schema-based document indexes and the Documents API.
- CrewAI — 1.15.23
Added native Gemini 3.8 Flash support and AMP evaluation of the last traced run.
- Consensus — September 28th, 2026 - Suggested Papers in My Library
Consensus now suggests papers based on what you’ve saved in My Library and your Collections.
- CodeRabbit — CLI v0.8.0 | CLI
The CLI adds deep reviews, included-review usage information, structured agent output, and cloud task handoff.
- Topaz Labs — Speed Update
Upscale video up to 10x faster; Topaz for Android debuts; SDR-to-HDR conversion gets added to the Adobe Premiere Panel.
- Ada AI Agent — Availability rules for web import sources
You can set availability rules on a web import source; pages it adds get those rules automatically.
- ThoughtSpot Analytics — Updating the value of a connection variable bound to a password field did not take effect on the connection.
Password variable updates are now correctly propagated to the connection on next use, without requiring a workaround.
- Keploy — v3.6.78
The proxy frames chunked HTTP bodies by chunk size and fixes two replay hangs.
- Keploy — v3.6.79
Compose stacks are brought up through the compose library instead of the Docker CLI.
- Tonic Structural — v1877
You can now use the TONIC_MAXIMUM_SESSION_LENGTH_IN_HOURS environment setting to automatically end user sessions that reach a maximum length; increased the parallelism.
- Claude AI — Claude Sonnet 5.5 launch
Launched Claude Sonnet 5.5, the second model in our Claude 5.5 family.