CreangelRequest a demo
IFINDIT
LakeHouseAuth IAM

IFINDIT PQRS CLASSIFIER

Organize incoming communications, identify key information and track each document to support review and processing.

Prepare the documents

Prepare the submissions received for review.

Classification with AI

Organize communications and extract key data to guide your management.

Follow each processing

Know the status of each document and see the information obtained.

View all featuresRequest a demo

Key capabilities

A first look at what you can do with this product.

  1. Document Load

    Collect communications in different formats to start your review.

  2. Standardisation with OCR

    Make scanned documents legible and take advantage of their content.

  3. Extraction of information

    Review each communication with key data extracted from its content.

  4. Classification of communications

    Organize trades and formalities according to the content and terminology of your process.

  5. Phased follow-up

    See at what stage each document is at and respond to errors in a timely manner.

  6. Search and audit

    Find communications and check your information and history when you need it.

Explore our capabilities

Select a category in the map to explore its features.

features of IFINDIT PQRS CLASSIFIER

Choose a category or search across all product features. Browse the results page by page.

132 features

The search includes all categories.

All features

DOCUMENT MANAGEMENT 19 features

Automatic updating of the list

DOCUMENT MANAGEMENT

Regularly update the list of documents to show processing status.

Alternative storage on file system

DOCUMENT MANAGEMENT

For environments without cloud storage server, documents can be saved directly into the server file system.

Secure private cloud storage

DOCUMENT MANAGEMENT

All documents loaded are saved to a private storage server (MinIO), equivalent to an enterprise hard drive on the internal network.

Drag and Drop Files

DOCUMENT MANAGEMENT

The user can take documents from his desktop or folders and drop them directly onto the screen to upload them, without needing to search manually.

Massive File Load

DOCUMENT MANAGEMENT

Allows you to select and load multiple documents in one operation.

Explicit confirmation of load

DOCUMENT MANAGEMENT

The system does not upload files automatically; it requires the user to press a confirmation button to start the transfer.

Downloading original documents

DOCUMENT MANAGEMENT

It allows you to download to the user's computer the original file that was uploaded to the system.

Confirmation dialog before deleting

DOCUMENT MANAGEMENT

Before deleting documents a warning message is displayed indicating the names of the files that will be deleted, requesting confirmation from the user.

Individual deletion of documents

DOCUMENT MANAGEMENT

Delete button in each row to delete a specific document, with prior confirmation to avoid errors.

Massive deletion of documents

DOCUMENT MANAGEMENT

When several documents are selected, a button appears that deletes them all simultaneously, with confirmation dialog.

Real-time update indicator

DOCUMENT MANAGEMENT

While the list is automatically refreshed, an animated icon tells the user that the data is being updated.

Paged list of documents

DOCUMENT MANAGEMENT

Displays the organization's documents in a paged table with number of configurable records.

Size limit per file

DOCUMENT MANAGEMENT

Supports files up to 50 MB.

Column Sorting

DOCUMENT MANAGEMENT

The user can sort the list of documents by name or by date of loading, by clicking on the header of each column.

Auto-renamed to duplicates

DOCUMENT MANAGEMENT

If a document with the same name already exists in the organization, the system automatically generates an alternative name by adding a number or suffix, avoiding overwriting...

File Selection from Explorer

DOCUMENT MANAGEMENT

Select button that opens the system file browser to choose one or more documents to load.

Multiple selection of documents

DOCUMENT MANAGEMENT

Check boxes in each row allow you to select several documents at once to apply batch actions to them.

Support for multiple file formats

DOCUMENT MANAGEMENT

The system accepts documents in PDF format, images (JPG, JPEG, PNG, GIF) and text documents (DOC, DOCX).

Preview files before confirming

DOCUMENT MANAGEMENT

Before sending documents to the system, a list of selected files with their name, size and type is displayed, allowing you to review and delete unwanted ones.

INTELIGENT DOCUMENT PROCESSING 16 features

Analysis by Apache Tika

INTELIGENT DOCUMENT PROCESSING

After OCR, the text is sent to a specialized service (Apache Tika) that extracts textual content more precisely from the processed PDF.

Detailed Classification of Crafts

INTELIGENT DOCUMENT PROCESSING

It classifies trades into categories such as requirements of control entities, reports and circulars.

Detailed Classification of Procedures

INTELIGENT DOCUMENT PROCESSING

It classifies procedures in legal categories, including exceptions, prescriptions, embargoes and payment facilities.

Classification using AI decision flows

INTELIGENT DOCUMENT PROCESSING

Apply AI decision flows to identify the type of document and execute the corresponding sorter.

Automatic OCR Need Detection

INTELIGENT DOCUMENT PROCESSING

The system automatically identifies which documents require text recognition (scanned) and which already have digital text, applying the process only where necessary.

Extraction of facts and foundations

INTELIGENT DOCUMENT PROCESSING

The AI analyzes the body of the document to summarize the arguments, reasons or facts that justify its existence or the contained request.

Extraction of the specific application

INTELIGENT DOCUMENT PROCESSING

Extracts the request or claim of the sender from the document.

Extraction of key legal terms

INTELIGENT DOCUMENT PROCESSING

The AI identifies and extracts the most relevant legal and administrative terms from the document, facilitating subsequent search and reference.

Extraction of the unique document identifier

INTELIGENT DOCUMENT PROCESSING

The system automatically detects and extracts the file/reference number or code of the document, if present.

Structured Metadata Extraction with AI

INTELIGENT DOCUMENT PROCESSING

Extract sender, recipient, subject matter, objective, summary, legal terms and date of receipt via AI.

Automatic document type identification

INTELIGENT DOCUMENT PROCESSING

Artificial intelligence determines whether a document is an Office (formal communication) or a Procedure (application requiring administrative resolution), classifying it before the...

Search engine indexation

INTELIGENT DOCUMENT PROCESSING

The content of each document, together with its extracted metadata, is indexed in Apache Soler to allow fast and efficient searches.

Interchangeable OCR motor

INTELIGENT DOCUMENT PROCESSING

The text recognition technology can be changed between different engines (OcrMyPDF, Tesseract) according to the needs of the organization, without modifying the rest of the system.

Standardisation of content (OCR)

INTELIGENT DOCUMENT PROCESSING

For scanned documents or images, the system automatically converts visual content into computer-readable text, allowing for further analysis.

Pipeline processing in three stages

INTELIGENT DOCUMENT PROCESSING

Each document follows a mandatory sequence: (1) Content standardization → (2) Metadata extraction → (3) Classification.

Spanish and English text recognition

INTELIGENT DOCUMENT PROCESSING

The optical character recognition engine processes documents in Spanish and English simultaneously, ideal for organizations with bilingual documentation.

FOLLOW-UP TO STATE AND TASKS 10 features

Animation of ongoing process

FOLLOW-UP TO STATE AND TASKS

When a document is being actively processed (normalizing, extracting or sorting), a rotating indicator appears on the corresponding chip, informing the...

Task Retry Control

FOLLOW-UP TO STATE AND TASKS

Tasks have a maximum number of configurable retry; if they fail repeatedly, they stop and record the error for revision.

Calculation of duration of tasks

FOLLOW-UP TO STATE AND TASKS

Each task records its start time, allowing to calculate how long it took to complete for performance analysis.

Full Life Cycle Task States

FOLLOW-UP TO STATE AND TASKS

The processing tasks pass through clearly defined states: Pending → In Execution → Successful / Failed, with record of each transition.

Visual process status indicators

FOLLOW-UP TO STATE AND TASKS

Each document in the list shows three indicators (N, E, C) that indicate whether it has already been Normalized, Extracted and Classified respectively, with colors that distinguish states...

Automatic temporary file cleaning

FOLLOW-UP TO STATE AND TASKS

After OCR processing, temporary files generated during analysis are automatically deleted from the server to free up space.

Time limit per task

FOLLOW-UP TO STATE AND TASKS

It applies time limits to standardisation and extraction tasks and cancels those that exceed them.

Detailed progress messages by stage

FOLLOW-UP TO STATE AND TASKS

During processing, the system updates descriptive messages in real time about what it is doing: getting the file, running OCR, sending to the engine of...

Prevention of duplicate tasks

FOLLOW-UP TO STATE AND TASKS

The system checks before starting any task if there is already an ongoing one for the same document and type of operation, avoiding processing the same file twice...

Registration of asynchronous tasks

FOLLOW-UP TO STATE AND TASKS

Each processing operation is recorded in the database as a task with unique identifier, type, status, action counter and progress messages.

VISUALIZATION OF METADATES 11 features

Details visibility condition

VISUALIZATION OF METADATES

The details panel is only enabled when the document has already been standardized and has its content extracted, avoiding displaying incomplete information.

Document Details Panel

VISUALIZATION OF METADATES

Expandable view within the table showing all metadata extracted from the document: category, subcategory, origin, recipient, subject, summary, objective...

Document audit view

VISUALIZATION OF METADATES

It shows who created the record in the system, who last modified it and the corresponding dates.

Category and subcategory view

VISUALIZATION OF METADATES

Displays the main classification (Office/Promitment) and the specific subcategory assigned by AI to each document.

Recipient view

VISUALIZATION OF METADATES

Shows who the document was addressed to and their position or organization.

Arrival date view

VISUALIZATION OF METADATES

It shows the date on which the document was received, as detected by AI in the content.

Facts and contentions

VISUALIZATION OF METADATES

Shows the arguments and reasons underlying the communication or request of the document.

View of specific request

VISUALIZATION OF METADATES

It clearly presents the specific request or claim contained in the document.

Source view and sender data

VISUALIZATION OF METADATES

It presents who sent the document, which organization it belongs to and its reference code or identifier.

Auto Summary View

VISUALIZATION OF METADATES

It presents a summary generated by AI with the most relevant points in the content of the document.

View of identified legal terms

VISUALIZATION OF METADATES

Lists the legal and administrative terms identified by AI in the document.

SEARCH AND INDIEXATION 7 features

Immediate update of the index

SEARCH AND INDIEXATION

Updates the index when a document is processed or changes its information.

Search with combined filters

SEARCH AND INDIEXATION

Searches can be combined with filters by specific fields (category, date, origin, etc.) for more accurate results.

Full Text Search

SEARCH AND INDIEXATION

It allows you to search for documents using words or phrases that appear anywhere in the document, not just in the title.

Deletion of the index when deleting documents

SEARCH AND INDIEXATION

When a document is removed from the system, it is also deleted from the search engine to maintain consistency between the two.

Apache Soler Enterprise Search Engine

SEARCH AND INDIEXATION

Index the content and metadata of documents in Apache Soler for consultation.

Search Pagination and Sorting

SEARCH AND INDIEXATION

Search results can be sorted by different fields and navigated on pages to handle large volumes of documents.

Verification of existence in the index

SEARCH AND INDIEXATION

Before updating or deleting records in the search engine, the system verifies that the document exists in the index, preventing errors.

ARCHITECTURE AND ASINCRON PROCESSING 9 features

Microservices architecture

ARCHITECTURE AND ASINCRON PROCESSING

The system is divided into independent components: web interface, API, processing workers, database, search engine, file storage and service...

Message queue with Redis

ARCHITECTURE AND ASINCRON PROCESSING

Link processing tasks to Redis.

Connection with automatic reconnection to the broker

ARCHITECTURE AND ASINCRON PROCESSING

Includes automatic reconnection to the messaging service when connection is interrupted.

Deployment with Docker containers

ARCHITECTURE AND ASINCRON PROCESSING

Pack the components in Docker containers for deployment and updating.

Background processing

ARCHITECTURE AND ASINCRON PROCESSING

Heavy tasks such as text recognition and AI sorting run in the background (with Celery), without blocking the user interface or affecting the user's...

Parallel document processing

ARCHITECTURE AND ASINCRON PROCESSING

It allows simultaneous processing of documents by asynchronous workers.

Inter-service internal network

ARCHITECTURE AND ASINCRON PROCESSING

The components of the system communicate with each other through a virtual private network (Docker network), without exposing internal services abroad.

Automatic reboot for failure

ARCHITECTURE AND ASINCRON PROCESSING

Sets the automatic restart of services for failure.

Separation between web server and workers

ARCHITECTURE AND ASINCRON PROCESSING

Separates the web server from document processing workers.

SAFETY AND AUTENTICATION 10 features

Authentication by token JWT

SAFETY AND AUTENTICATION

Access to the system is controlled by security tokens (JWT) that verify the user's identity in each operation, without the need to send passwords repeatedly.

Authorization heads in all applications

SAFETY AND AUTENTICATION

Each request from frontend to backend automatically includes the security token in HTTP headers.

CORS Configuration

SAFETY AND AUTENTICATION

Sets the authorized origins for web requests using CORS.

Control of permits by group

SAFETY AND AUTENTICATION

Users belong to groups with specific roles that determine what actions they can perform within the system.

Permission management from Redux

SAFETY AND AUTENTICATION

User permissions are loaded when logging in and stored in the application's global state, being available throughout the interface without repeated queries to...

Session information in the token

SAFETY AND AUTENTICATION

Each access token contains the user's identity, organization, email, username and session identifier, allowing precise authorization.

Integration with Creangel authentication system

SAFETY AND AUTENTICATION

The system connects to Creangel Technologies' central authentication platform to verify users and obtain their permissions and groups.

Restriction of permitted file types

SAFETY AND AUTENTICATION

The system only accepts file formats defined in its configuration, rejecting any files of type not allowed.

Maximum file size validation

SAFETY AND AUTENTICATION

The system verifies that no file exceeds the limit of 50 MB before processing.

Environment variables for credentials

SAFETY AND AUTENTICATION

Manages the credentials of services using environment variables.

USER INTERFACE 13 features

Side Panel File Search Bar

USER INTERFACE

In the loading panel, the user can filter the selected files to upload by typing part of the name.

Catching undefined routes

USER INTERFACE

Any URL that does not correspond to a known page is captured and redirected to the corresponding error screen.

Load indicator in document table

USER INTERFACE

While the table data is loaded, a progress indicator is displayed so that the user knows that the system is working.

Handling of pages not found (404)

USER INTERFACE

When a user accesses a non-existent address, a friendly page is displayed indicating that the path was not found, with option to return.

Visual handling of system errors

USER INTERFACE

If an unexpected error occurs in the application, a friendly error screen is displayed instead of a blank screen or technical message.

Context menu of actions by document

USER INTERFACE

A drop-down menu in each row shows the actions available according to the status of the document (Normalize, Extract, Sort), automatically hiding the actions already...

Navigation of pages with position indicator

USER INTERFACE

Previous/next page buttons with counter showing the range of current records (e.g. 1-10 out of 45), facilitating orientation in large listings.

Specific error notifications

USER INTERFACE

When something fails, the system displays a descriptive message indicating which operation failed and in which document, helping the user understand the problem.

User-configurable rowing

USER INTERFACE

The user can change how many documents are displayed per page by selecting between predefined options (5, 10, 50, 100, 500).

Side panel of persistent loading

USER INTERFACE

Keeps the loading panel available while the user consults the document list.

Project information page

USER INTERFACE

Dedicated section with information about the project, its team and organizational context.

Home page with system description

USER INTERFACE

The main page presents the IFINDIT system, its purpose and main functionalities visually.

On-screen notification system

USER INTERFACE

Informative messages emerge on the screen informing the result of each operation (successful loading, error in sorting, process initiated, etc.), disappearing automatically...

INTEGRATION WITH ARTIFICIAL INTELIGENCE 8 features

Sorting with structured output

INTEGRATION WITH ARTIFICIAL INTELIGENCE

Request the AI structured results in JSON for processing by the system.

Classification by function with invocation of tools

INTEGRATION WITH ARTIFICIAL INTELIGENCE

Use invocation of tools to request structured classifications.

Decision Graphs with LangGraph

INTEGRATION WITH ARTIFICIAL INTELIGENCE

The classification flow is modeled as a graph of states with LangGraph, allowing conditional ramifications (Office vs.

Integration with OpenAI GPT-4o Mini

INTEGRATION WITH ARTIFICIAL INTELIGENCE

The system uses OpenAI's GPT-4o Mini language model to perform metadata extraction and document classification, taking advantage of its data processing capabilities.

Managing errors in AI responses

INTEGRATION WITH ARTIFICIAL INTELIGENCE

If the AI does not return a valid response or fails, the system captures the error and continues with a default value, avoiding process interruption.

Prompts specialized in Colombian legal language

INTEGRATION WITH ARTIFICIAL INTELIGENCE

It uses classification instructions with definitions of Colombian documents and legal terminology.

Zero temperature for consistency in classification

INTEGRATION WITH ARTIFICIAL INTELIGENCE

Set the model to zero temperature to reduce variation in classifications.

Use of LangChain for AI orchestration

INTEGRATION WITH ARTIFICIAL INTELIGENCE

LangChain is used as an abstraction layer to build conversation flows with AI, facilitating system maintenance and evolution.

DATA MANAGEMENT AND DATA BASE 9 features

PostgreSQL relational database

DATA MANAGEMENT AND DATA BASE

All system information (documents, metadata, tasks) is stored in a PostgreSQL database.

Metadata Cascading

DATA MANAGEMENT AND DATA BASE

When deleting a document, its associated metadata are automatically deleted, maintaining the referential integrity of the database.

Advanced List Filtering

DATA MANAGEMENT AND DATA BASE

The API supports complex filters in document listings, including text search in specific fields, sorting and pagination.

Document metadata in separate table

DATA MANAGEMENT AND DATA BASE

The information extracted by AI (category, subcategory, summary, terms, etc.) is stored in a separate table linked to the document, maintaining the data model...

Automated database migrations

DATA MANAGEMENT AND DATA BASE

Changes in the database structure are managed by Alembic, allowing the schema to be updated in a controlled and versioned manner.

Model document with processing statuses

DATA MANAGEMENT AND DATA BASE

Each document records its status in the pipeline: if it requires OCR, if its content was standardized, if it is under normalisation, if it was extracted, if it is under extraction, if it was...

Audit log by registry

DATA MANAGEMENT AND DATA BASE

Each record in the database saves who created it, who last modified it, and the creation and editing dates.

Restriction of oneness by name and organization

DATA MANAGEMENT AND DATA BASE

The system allows that two documents with the same name cannot exist within the same organization, preventing duplicates.

Support for database text arrangements

DATA MANAGEMENT AND DATA BASE

The legal terms extracted are stored as text arrangements in PostgreSQL, allowing multiple values in a single field.

REST API AND WEBSITE SERVICES 13 features

REST API versioned

REST API AND WEBSITE SERVICES

All system services are available via an REST API under the /v1/ path, facilitating maintenance and compatibility with future versions.

Document Update Endpoint (PUT)

REST API AND WEBSITE SERVICES

Allows you to change the name of an existing document.

Document sorting endpoint

REST API AND WEBSITE SERVICES

Starts the IA sorting process for a specific document, returning a task ticket.

Document Consultation Endpoint (GET)

REST API AND WEBSITE SERVICES

It allows you to obtain the detailed information of a specific document using its unique identifier.

Metadata Query Endpoint

REST API AND WEBSITE SERVICES

It allows you to obtain all metadata extracted from a document (category, subcategory, summary, legal terms, etc.).

Document Creation Endpoint (POST)

REST API AND WEBSITE SERVICES

It allows you to upload one or more files to the system in a single request, creating corresponding records in the database.

Original File Download Endpoint

REST API AND WEBSITE SERVICES

Allows downloading the original document file as it was loaded.

Document Removal Endpoint (DELETE)

REST API AND WEBSITE SERVICES

Removes a document from the system, including its file in storage, its record in the search engine and all its metadata.

Endpoint extracting content

REST API AND WEBSITE SERVICES

Starts the process of extracting metadata with AI for a specific document, returning a task ticket.

Listing Endpoint with Advanced Parameters (POST)

REST API AND WEBSITE SERVICES

It allows you to obtain lists of documents with pagination, sorting, filtering and total counting options.

Content Standardisation Endpoint

REST API AND WEBSITE SERVICES

Start the OCR process for a specific background document, returning a task ticket for tracking.

Standardized API Responses

REST API AND WEBSITE SERVICES

All API responses follow a uniform format with status, message and data fields, making it easier to integrate with other applications.

Apache Tika Text Extraction Service

REST API AND WEBSITE SERVICES

Independent Apache-based Tika microservice that receives PDF files and returns textual content; supports PDF, DOC, DOCX and TXT.

STATE MANAGEMENT (FRONTEND) 7 features

Cola of notifications in global status

STATE MANAGEMENT (FRONTEND)

Notifications of success or error are managed through the global state.

Overall implementation status with Redux

STATE MANAGEMENT (FRONTEND)

The web application uses Redux to manage a centralized state.

Management of enabled actions

STATE MANAGEMENT (FRONTEND)

The system dynamically controls which actions are available to the user according to the current state of the application.

Information management of the organization

STATE MANAGEMENT (FRONTEND)

The data of the organization to which the user belongs are loaded on startup and kept available throughout the session.

User information management in session

STATE MANAGEMENT (FRONTEND)

The name, identifier, mail and other data of the connected user are stored in the global state and are available throughout the application.

User permission management

STATE MANAGEMENT (FRONTEND)

The user's groups and roles are stored in the global state, allowing to display or hide functionalities according to permissions.

Sessions Persistence in Storage

STATE MANAGEMENT (FRONTEND)

The user session status is saved in the browser storage to keep the session active if the user reloads the page.

Part of the IFINDIT platform

Explore how it relates to other products and shared capabilities.

Persons and equipment

Consultation · Analysis · Information management

Sources of information

Documents and Repositories · Databases · Applications and APIs

A functional view of IFINDIT and its shared capabilities. Each product connects different needs of the organization.

View platform architecture (SVG)

Let's talk about your information.

Tell us what you need to find, analyze or organize. Discover IFINDIT with the Creangel team.

Request a demoExisting customer? Get support

Image