Tools for Seamless External Integration

We are excited to introduce Tools, a powerful new capability in Lamatic AI that enhances automation and expands the system's ability to interact with external functions and APIs. Tools act as a structured bridge between natural language understanding and programmatic execution, allowing AI to interpret user requests and seamlessly trigger specific actions.

🔹 Key Benefits:

• Seamless API & Function Integration – Connect AI models with external services and execute complex workflows.

• Enhanced Automation – Reduce manual effort by automating tasks through AI-driven actions.

• Improved Efficiency – Streamline operations by enabling direct execution of commands via natural language.

• Expanded AI Capabilities – Go beyond text-based interactions with real-world task execution.

Checkout the docs here 

Model Parameters Support

We have introduced Model Parameters Support, allowing users to fine-tune AI behavior with greater control over model outputs. This feature provides customizable settings such as temperature, frequency penalty, presence penalty, top-p sampling, max tokens, and more.

🔹 Key Benefits:

• Fine-Tuned AI Responses – Adjust temperature and sampling parameters for more deterministic or creative outputs.

• Customizable Model Behavior – Modify penalties, biasing, and stop sequences for tailored responses.

• Optimized Performance – Control token limits to balance response length and processing efficiency.

Checkout the docs here

Other Improvement

1. Enhanced API Node to support multiple REST API integrations for smoother automation and data exchange.

2. Improved Zod Schema Builder for a more intuitive and user-friendly experience.

3. Simplified schema creation, validation, and management for developers.

IBM watsonx models

Introduced IBM Watsonx as a Model Provider. Tailored for enterprises, Watsonx ensures scalability, security, and reliability for integrating AI into workflows.

These model includes: Mixtral, Allam, Granite, Mistral, LLaMA, Pixtral, Flan, Jais

Check out the documentation here

Extract from File Node Update

1. Enhanced the "Extract from File" node for improved performance.

2. Automatically detects and supports various file formats.

3. Provides a consistent output structure for all supported file types.

4. Check out the docs here

AWS S3 Node Update

1. Improved functionality of the AWS S3 Node.

2. Added document_url output, allowing seamless integration with the "Extract from File" node for data extraction.

3. Explore the use cases of the AWS S3 Node: Click Here

Bug Bounty Program

At Lamatic, we value collaboration with the global security research community to strengthen our product security.

1. Researchers are invited to identify vulnerabilities in our systems.

2. Rewards are provided for valuable contributions.

3. For more details, visit: Bug Bounty Program

Learning Guides: Boost Your AI Development

Our Learning Guides are continually updated to help developers harness the full potential of Lamatic.ai. Step-by-step instructions for implementing advanced AI features.

🎓 Start exploring: Lamatic.ai Learning Guides

Other Improvement

1. Enhanced Studio UI for better accessibility and performance.

2. Updated documentation with improved details for all nodes.

Introducing the Extract from File Feature

We’re excited to announce the launch of the Extract from File feature! This versatile tool is designed to simplify data extraction from various file formats, offering a flexible and customizable interface. With its robust configuration options, you can accurately parse and format data for seamless downstream processing. Whether you’re working with documents, spreadsheets, or other file types, the Extract from File node ensures precision and efficiency, empowering you to streamline your workflows like never before.

Benefits

1. Versatility: Single node solution for handling various file formats commonly used in data processing.

2. Precision Control: Fine-grained control over data extraction through format-specific configuration options.

3. Data Quality: Built-in options for data cleaning and validation during extraction.

4. Seamless Integration: Easy integration with other nodes in the workflow for comprehensive data processing.

📄 Learn more in our documentation: Extract from File Node

Other Improvements

1. Enhanced Studio UI for Flow Management

   • We’ve revamped the Studio UI to make managing flows more intuitive and efficient.

   • The interface now supports dynamic titles for flows, improving clarity and enabling a more personalized experience when working with your projects.

Seamless S3 Integration: Build Advanced Data Workflows with Ease

We’re excited to introduce a powerful new feature on the Lamatic.ai platform—a dedicated AWS S3 Node for seamless data collection from Amazon S3 buckets.

AWS S3 Node Features

The AWS S3 Node simplifies and automates the process of fetching and synchronizing files from your S3 buckets. It supports a variety of file types, including:

• Text files

• PDFs

• Word documents

• Other formats stored in the bucket

This node is designed to facilitate regular file synchronization, preparing your data for vectorization and indexing to power Retrieval-Augmented Generation (RAG) workflows.

Key Benefits

• Streamlined File Collection: Automates the retrieval of files from S3 buckets, saving time and effort.

• Optimized for RAG Flows: Prepares files for vectorization and indexing to enhance AI-driven retrieval and generation.

• Scalability: Efficiently handles growing data volumes, ensuring smooth scaling as your needs evolve.

📄 Learn more in our documentation: AWS S3 Node Docs

Smart Model Recommendations

We’re excited to make choosing the right model for your use case effortless with Smart Model Recommendations.

🔍 Here’s how we guide you:

• Preferred: The best balance of speed and quality—our go-to recommendation.

• Best Quality: Top-tier response quality for when precision matters most.

• Fastest: Blazing-fast response times for real-time needs.

• Cheapest: The most cost-effective model for budget-conscious projects (includes input + output token costs).

💡 Why this matters:

• Choosing the right model has always been tricky.

• With Lamatic, it’s now effortless.

• Whether you value speed, quality, or cost, we’ve got you covered.

📄 Explore the details in our documentation: Model Recommendations

Learning Guides: Boost Your AI Development

We’re also introducing Learning Guides to help you get the most out of Lamatic.ai. These guides provide step-by-step instructions to help developers implement advanced AI features with ease.

🎓 Start exploring: Lamatic.ai Learning Guides

Other Improvements

1. Enhanced Studio UI for Flow Management

   • We’ve revamped the Studio UI to make managing flows more intuitive and efficient.

   • The interface now supports dynamic titles for flows, improving clarity and enabling a more personalized experience when working with your projects.

2. UI Improvements Across RAG, Index, and Search Nodes

   💡 Benefits: Intuitive UI, fewer input errors, improved discoverability, and streamlined setup.

   1. RAG Node Filters:

      • Replaced Monaco JSON input with a schema-based interface for improved usability.

      • New UI mimics Zod schema structure for better clarity and validation.

   2. Index Node Primary Keys:

      • Added a dropdown menu for available properties with auto-suggestions.

      • Improved discoverability of valid primary key options.

   3. Numeric Controls:

      • Introduced slider controls for range-based values, including a visual slider for the certainty parameter.

      • Enhanced UX for numerical input fields.

   4. Search Node Enhancements:

      • Added new configuration options: Autocut parameter, weights configuration, and additional search refinement controls.

Adding 10 new model providers → Expanding Model Capabilities

We have upgraded our model proxy system to support several new models, including:

• Cloudflare

• Cohere

• xAI

• AWS Bedrock

• Azure

• Cohere

• Perplexity

• FireworksAI

• DeepInfra

• Together AI

In addition, we now provide recommendations for the optimal model from each provider, categorized as follows:

• Preferred: The default model we recommend, offering the best balance of speed and quality.

• Best Quality: The model with the highest overall response quality.

• Fastest: The model with the quickest response time.

• Cheapest: The model with the lowest token cost (including both input and output).

Check out the doc: https://lamatic.ai/docs/models

Execute Flow within a Flow

Agentic Systems now support complex workflows where you can invoke other Agentic Flows to achieve your desired outcomes. This new capability allows for the creation of intricate, multi-step processes. Simply select the Execute Flow node from the node selector and choose the Flow you want to run.

Currently, you can select flows that use Lamatic-built interfaces as trigger nodes, such as GraphQL, Chat, and Search.

Check out the Doc: https://lamatic.ai/docs/node/execute-flow

Other Improvements

Improved Support for Role based Access Controls

We have significantly improved support for Role-Based Access Control (RBAC) to ensure a more secure, scalable, and flexible management of permissions within the system.

Easier Integration of third party applications

We have simplified the process of integrating third-party applications into your flows, making it more efficient and user-friendly.

Passing context history in JSON Generator

The JSON Generator now supports passing context history, enabling more dynamic and context-aware responses. This feature allows you to include historical data or previous interactions as part of the JSON structure to improve the flow and relevance of your output.

Web Search Node

Introducing Web Search Nodes – empowering you to seamlessly search the internet and incorporate real-time data into your agentic flows.

Highlights -

• Easy Setup: Configure your Web Search Node with Serper API credentials and pass your query.

• Comprehensive Results: On flow execution, the node fetches relevant search results you can use downstream.

• Filtering Capabilities:

• Search Types: News, Images, Videos, Scholar, Places, Shopping.

  • Date Range: Fetch results from specific timeframes.

  • Location & Language: Narrow down searches by country and preferred language.

👇🏻 Checkout the docs for more information and start using it for free.

Web Search Documentation

Improved Logs

Our logs have been completely rebuilt to help you monitor, analyze, and export data effortlessly.

Highlights:

• Improved UI: A minimalist design that displays critical information at a glance.

• Token Usage: Easily track the tokens consumed per execution.

• Export Options: Download logs as CSV or JSON for deeper analysis.

👇🏻 Checkout the docs for more information.

Logs: Monitoring and Analyzing Requests in Real Time

Other Improvement

1. Default Models: Set default models for seamless configuration which auto-picks the model on any new AI Node. Changing default model changes it everywhere you chose default model.

2. Job Scheduling: Simplified scheduled options to setup batch processes.

3. Save Before Exiting: Preserve dialogues before leaving a flow.

4. Variable Selector: Improved the variable insertion menu to easily pass info from one node to another.

5. Debugger Enhancements: Fixed and tweaked testing of flows and nodes.

6. Project Management:Delete or pause projects effortlessly.

7. Delete Vector Stores when no longer needed.

🎨 New Flow Editor

• Intuitive Design: Completely revamped for simplicity. Now, you can add nodes with a single click, and instead of connecting lines, you can seamlessly arrange nodes in designated areas for a smoother, cleaner workflow.

• Enhanced Usability: No more tedious dragging! The improved layout makes complex flows feel intuitive and easy to navigate.

• Streamlined Experience: Editing, arranging, and organizing your workflow has never been more efficient.

Read Docs

🏠 New Home View

• Modern and Elegant Layout: The new home view provides a visually pleasing, streamlined overview of your project.

• Quick Access to Information: Get all relevant project details at a glance, making it easier to monitor progress and access key components.

📄 Templates

• Pre-built Flow Templates: Start fast and efficiently with pre-configured templates tailored for various use cases.

• Customizable Options: Easily adapt each template to suit your unique requirements, cutting down on setup time and getting you to your goals quicker.

Read Docs

💬 ChatWidgets

• Embedded Chat Support: Add a fully functional chatbot to your website, enhancing user engagement and customer service.

• Customizable and Scalable: Tailor the chatbot experience to meet the needs of your audience, from simple Q&A to complex interactions.

Read Docs

🔗 N8N Integration

• Connect with 1000+ Apps: Access the powerful automation capabilities of N8N to integrate with numerous applications seamlessly.

• Enhanced Automation: Build workflows that bring together various tools, increasing efficiency and minimizing manual intervention.

⚡️ Additional Features

🤖 Agent Classifier

• Intent Detection: Automatically classify and route user intents, allowing your flows to follow the most relevant path for each input.

• Improved Accuracy: The classifier ensures that interactions are contextually accurate, enhancing user satisfaction and flow efficiency.

🌐 Webhooks

• Seamless Integration: Trigger flows directly from third-party sources, enabling dynamic responses and actions in real-time.

• Asynchronous Requests: Execute tasks asynchronously for faster and more flexible performance.

Read Docs

🔮 Multimodal Generator

With Multimodal Node, you have the capability to enhance your interactions with language learning models (LLMs) by including media content alongside your textual prompts. This innovative feature allows the LLMs to not only process the text you provide but also to "see" and contextualize media.

🛠️ Build Flows in Low Code YAML

We are introducing a new way to build and debug workflows. We understand that sometimes the UI can be limiting in terms of selecting options, and you may want to manually override and take control of those parameters. We are introducing a new low-code YAML configuration, which provides you with a mirrored output of the workflow and syncs back changes to the workflow when you update it. You can also copy node blocks or even entire workflows to recreate the same functionality.

C

heckout the Docs

🔍 Hybrid & Keyword Search Node

We are extending the vector search functionality and adding 2 new nodes for keyword and hybrid Search. Keyword search, also called "BM25 (Best match 25)" or "sparse vector" search, returns objects that have the highest BM25F scores. Hybrid search combines the results of a vector search and a keyword (BM25F) search by fusing the two result sets. With now different search method available, you can build any type of search experience on your App.

📋 Log Search and Filtering

You can now search and filter logs to find specific events, errors, or patterns more efficiently, allowing for quicker troubleshooting and analysis of system performance. This enhanced capability improves overall system reliability and helps in proactively identifying potential issues before they escalate.

🚦 Status Monitor

Lamatic.ai status

You can now see the real time status of all the Lamatic.ai services along with the response time.

status.lamatic.ai

⚡️ Additional Features

Support for history context in RAG, Text Generator and Multimodal node

To make it easier for developers to adapt and integrate the Lamatic GraphQL API, we have added an automatic API docs generator right inside the project. The docs are created in Python and JavaScript, along with a prefilled GraphiQL playground. All developers need to do now is copy and paste.

Create and update test from Flow editor

Create new test and update existing test cases for your workflow directly inside the editor. This allows for seamless integration and ensures that your testing process is efficient and up-to-date, enhancing the overall quality and reliability of your workflow.

SignIn with Github and Google

You can now add credentials for apps directly from the workflow, eliminating the need to switch back and forth between the App sections.

🛠️ Bug Fixes & Improvements

• Improve: Improved Model Credential Experience

• Optimize: Vector Indexing

• Improve: Fetch First Row of Google Sheets

• Improve: Navigation and automatic Project Routing to Last active

• Fixed: Issues with jobs not showing correctly

• Improved: Add project Slug.

• Improved: Auto Add schema matching payload to test cases.