​

New LLMs: Claude 4 Series

Added Anthropic’s latest flagship models:

• Claude 4 Sonnet - Enhanced performance and reasoning capabilities
• Claude 4 Opus - Most advanced reasoning model with superior problem-solving abilities

​

New LLM: Gemini Flash 2.5

Added Google’s latest Flash model, delivering efficient and lightning-fast responses for high-throughput applications.

​

New LLM: OpenAI o3

Added OpenAI’s latest general availability model, now available for all users.

​

New LLM: GPT 4.1

Added OpenAI’s newest GPT model, featuring enhanced reasoning capabilities and improved performance across all task types.

​

New LLM: Gemini Pro 2.5

Added Google’s latest Pro model. Features include:

• Advanced reasoning capabilities
• Enhanced coding assistance
• Improved multimodal understanding

​

Knowledge Base Foldering and Management

Major update to the Knowledge Base system:

• Introduced folder organization with agent-specific enable/disable controls
• Significantly improved loading performance for agent chat interface and knowledge base
• Added bulk operations for knowledge base entries:
  • Mass move functionality
  • Batch deletion
  • Bulk enable/disable for agents

​

Agent Builder Templates

Enhanced agent creation experience with new startup options:

• Template-based agent creation
• Start from scratch option
• Integrated tutorial video guidance for the Agent Builder

​

New LLM: GPT-4.5

• Added OpenAI’s latest non-reasoning LLM.

​

Landing Page Update

• Refreshed landing page with improved user experience and updated content.

​

Pipedrive CRM Integration

Added 21 new agentic tools for Pipedrive CRM.

​

Builder Interface Segmentation

Reorganized the Builder interface into three distinct sections:

• Core: Essential components including system message, tools, and triggers
• Cosmetic: Visual and non-functional customization options
• Advanced: Power user features like variables and preferred LLM settings

​

Notification Improvements

• Disabled email notifications for batch mode operations with Human in the Loop tools

​

LLM Selection Enhancement

• LLM selector now properly affects request regeneration

​

Bug Fixes

• Resolved Human in the Loop functionality issues
• Fixed the Agent Builder agent related bugs

​

New LLM: Claude 3.7 Sonnet

• Added Anthropic’s latest hybrid reasoning model.
• Now designated as our “Smartest” category model.

​

Apollo Integration

Added Apollo company enrichment tool. Agents can now:

• Look up company data (revenue, employee count, etc.)
• Access Apollo’s business database directly

​

Batch Mode Monitoring

• Now, you easily monitor the reasoning process for each row of the batch operation, which gets a separate branch within a thread.

​

Re-generate

• Easily re-generate any response with a single click

​

Branching Conversations

• You can now Edit your previous inputs to generate a new response in a new branch.
• If you want to make an edit with a new LLM, simply change it in the LLM selector and save the edits.

​

New LLM: o3-mini

o3-mini is OpenAI’s most recent small reasoning model, providing high intelligence at the same cost and latency targets of O1-mini.

​

Token Counter

• You can now view the number of tokens that the current chat had consumed in the Chat UI

​

Bug Fixes

• Human in the loop triggering for tools where HiLT is not turned on fixed
• LLM selector resetting to preferred LLM after every request
• Mobile compatibility and accessibility fixes
• and more…

​

New LLMs by DeepSeek

Added DeepSeek’s newest LLMs:

• R1: (open-source reasoning specialist) + V3 (enterprise-grade inference)

New OpenAI LLM:

• Introduced ChatGPT-4o-latest (auto-updating to OpenAI latest ChatGPT version)

​

Agent Configuration Limits

Set maximum configuration parameters:

• 10 tools per agent
• 10 collaborators per agent

​

Human-in-the-Loop for tools

Added tool execution approval controls:

• Per-tool approval settings in Builder
• Add custom reasons when rejecting actions
• Get notified through email and sidebar indicators

​

Pricing Model Restructure

• Introduced flexible credit bundles for Pro/Business tiers
• Business plans now include 10 default user seats

​

Stripe Integration & Tools

Added 5 new Stripe tools including:

• Charge refund management
• Subscription lifecycle controls
• Customer/subscription search capabilities

​

Bye Compute Points!

Renamed “compute points” to “credits” across all platform interfaces.

​

Free Tier Enhancement

Increased Knowledge Base entry limit from 5 to 50 entries for free accounts.

​

Data Analysis

Added CSV file upload capability to the Data Analysis Tool for streamlined data import processing.

​

Search Integration

Implemented direct Google Search integration for real-time web query access.

​

Output Formatting

JSON outputs now display in formatted, hierarchical structure for improved readability.

• Added proper indentation and line breaks
• Implemented syntax highlighting
• Standardized formatting across all output displays

​

Analytics

Added usage statistics chart of your AI Agents to the analytics dashboard.

​

Search and Filter

Added thread search functionality and improved sidebar organization for better content discovery.

• New search dialog enables quick thread location. ⌘/Ctrl + K to open dialog
• Sidebar now supports filtering threads by AI agent and include external users threads. ⌘/Ctrl + J to open dialog
• Redesigned sidebar layout to accommodate new features

​

AI Vision Updates

Fixed image processing triggers and improved source validation.

• Eliminated false triggers from irrelevant links
• Restricted Vision activation to system-provided image sources only

​

Builder Improvements

Added underscore character support in variable naming conventions.

​

Grok-2 Model Integration

Added xAI’s Grok-2 text model to the platform’s available language models. The integration supports structured outputs and enhanced processing capabilities.

​

Analytics Interface Update

Reworked a Analytics dashboard that consolidates key metrics:

• Compute Point usage tracking
• Daily usage monitoring
• AI agent Statistics

​

Technical Improvements

• Fixed date range selector display issues
• Resolved AI agent edit page navigation errors
• Added direct memory feature toggle controls in dialog interface

​

Google Sheets Tools

Added comprehensive Google Sheets functionality with tools for spreadsheet management:

• Create new spreadsheets
• Add and update row data
• List spreadsheets, tabs, and row contents
• Search files
• Modify row properties

​

Data Analysis Tool

Added Python-based data analysis capabilities for processing and visualizing data. Powered by E2B.

​

Improved Tool Management

Implemented categorized tool organization system using checkmark tree structure for improved tool discovery and addition.

​

Notion, Mobile and other bug fixes

Resolved issues with Notion tools functionality, improved mobile compatibility, and fixed authentication page redirect behavior.

​

Memory System

New memory management system enables agents to store and utilize information automatically. Added “Memory” tool required for functionality.

• Automatic storage of relevant facts
• Ability to edit existing memories

​

Variables System

Added variable placeholders in system messages that end users can customize without editing the core agent configuration.

• Simplifies agent reuse across different contexts
• Enables template-based agent use / cloning

​

Subscription Management

Added subscription portal for account management and billing control.

• Self-service subscription modifications
• Affiliate program with promo code support

​

Agent Builder Performance

Improved builder interface performance with reduced latency.

​

LLM Categories Introduced

Introduced three dynamic LLM categories that our team updates as new LLMs become available.

• Auto: Balanced performance model optimized for general use and stability
• Smartest: Always uses the highest-performing available LLM, automatically updating when superior models are released
• Efficient: Cost-optimized model that maintains smart performance while minimizing resource usage

​

Token Count Fixed

Token count for all LLMs has been fixed.

​

Bug Fixes

• Pagination has been added to the Knowledge Base and other tables
• A bug preventing agent creation/editing has been resolved
• Fixes for issues related to conversation started deletion
• Improvements made to organization invitation functionality

​

Notion Tools Added

• Look up databases.
• Add an item to a database.

​

Video Caption Scraper Tool

• You can now scrape available captions from videos from YouTube, X, and other social media.

​

Email Triggers Modified

Email triggers have been updated to include document contents.

​

Twitter Tools

New tools introduced for interacting with Twitter data.

• Tool for scraping information and tweets from specific Twitter profiles
• Tool for retrieving the latest or top tweets for specific keywords

​

LinkedIn Tool

A new tool allows agents to scrape profile information from LinkedIn by URL.

​

Intercom Tool and Trigger

New Intercom tool and trigger:

• New incoming Intercom messages can now trigger agents
• Agents now have a tool to send an Intercom message

​

Vision Capabilities

Agents can now accept images as input.

​

Weclome Screen Re-design

The welcome screen has been redesigned to accommodate longer welcome messages and instructions.

​

Bug Fixes

Several issues have been resolved.

• Older threads not being loaded fixed
• Bug fixes related to workspaces
• Time trigger based bugs fixed
• Various smaller fixes

​

Enhanced AI Teammate Interactions

We’ve significantly improved how AI Teammates communicate:

• AI Teammate messages now appear as separate, chronologically ordered entries.
• Each message includes details about tools used and all relevant information related to the request.
• Tool usage is now displayed in the exact order and context of the agent’s reasoning.

These updates provide greater transparency into AI Agent decision-making and collaboration processes.

​

Revamped HubSpot Integration

Our HubSpot tools now offer more flexibility and faster performance:

• Search for deals by name, in addition to finding them by ID.
• Significantly reduced trigger and message sending times for HubSpot-related tasks.

​

Global App Improvements and Redesign

Invicta AI has received a comprehensive update:

• Enhanced user experience with intuitive navigation and smoother interactions.
• Introduced a new Explore Page to discover AI Agent capabilities more easily.
• Implemented various UX improvements throughout the platform.

​

New Landing Page

Visit our updated website to explore:

• Detailed use cases to help you understand how AI Agents can benefit your business.
• New tutorials to get you started quickly.
• Expanded information for a comprehensive overview of our platform.

​

Revamped Organizations & Workspaces

We’ve restructured how teams collaborate on Invicta AI:

• Introduced organization-level credits for better resource management.
• Added admin capabilities for enhanced control over your organization’s AI Agents.
• Implemented email invitations to easily onboard team members to your organization.

​

New Agent Builder

Create custom AI Agents tailored to your specific needs through an intuitive chat interface that guides you through the agent creation process.

​

Expanded Documentation

We’ve significantly enhanced our documentation to help you get the most out of Invicta AI:

• Comprehensive guides on all agent components.
• Detailed explanations of available tools and their schemas.
• In-depth tutorials on advanced features and best practices.

​

Login Issues Resolved

We’ve addressed caching-related login problems to ensure a smooth and reliable user experience.

​

New Use Cases Page

Explore real-world applications of Invicta AI in action, including customer support scenarios and scheduling use cases.

​

Clone Team Feature

Easily replicate entire AI Agent teams with a single click:

• Clone individual agents or entire teams, including all downstream AI Teammates.
• Maintain complex agent relationships and hierarchies when duplicating.
• Quickly scale your AI workforce by replicating successful team structures.

This feature dramatically reduces the time needed to set up multiple similar AI Agent teams.

​

Airtable Integration

Empower your AI Agents with Airtable capabilities:

• Import data from Airtable into your AI workflows.
• Edit and update Airtable records through AI Agent actions.
• Delete entries when necessary, all managed by your AI Agents.

This integration allows for seamless data management between Invicta AI and your Airtable bases.

​

Outlook Calendar Tools

Enhance your scheduling capabilities with Outlook Calendar integration:

• AI Agents can now view and import calendar events.
• Create new events directly from AI conversations.
• Manage your Outlook Calendar effortlessly through AI interactions.

These tools streamline your scheduling processes and keep your calendar up-to-date.

​

Google Calendar Integration

Similar to Outlook, we’ve added robust Google Calendar support:

• Import and view Google Calendar events within AI conversations.
• Create new events based on AI Agent suggestions or user requests.
• Seamlessly manage your Google Calendar through Invicta AI.

This integration ensures your AI Agents are always in sync with your Google Calendar.

​

Enhanced Outlook Tools and Triggers

We’ve significantly improved our Outlook email integration:

• AI Agents can now send and reply to emails directly.
• Support for CC and BCC functionality in email communications.
• Implement email-based triggers for automated AI Agent actions.

These enhancements make email management through Invicta AI more powerful and flexible.

​

Gmail Integration

We’ve added comprehensive Gmail support to Invicta AI:

• Send and reply to emails using AI Agents.
• Utilize CC and BCC features for more complex email workflows.
• Set up Gmail-based triggers for automated AI responses and actions.

This integration brings the power of Invicta AI to your Gmail inbox.

​

New Image Generation Tools

Expand your AI Agents’ capabilities with cutting-edge image generation:

• DALL-E 3 integration for high-quality, AI-generated images.
• Flux (Pro, Dev, Schnell) support for diverse image creation options.
• Empower your AI Agents to create and manipulate images as part of their tasks.

These new tools open up exciting possibilities for visual content creation within your AI workflows.