Connecting Devices to the LIMS Using the Import Interface Editor

Import Interfaces: Automatically Import Data into the LIMS

Connect analytical instruments, testing machines, third-party laboratories, and ERP systems such as SAP to the LIMS: CSV, JSON, XML, REST API, and SFTP—with AI-powered mapping.

Everyday Life Without an Interface

Four Things That Take Time in the Lab — and Lead to Errors

Values are entered manually

The device prints or exports the data, and someone enters the numbers into the system by hand. Each value is handled twice, and that's when the transposed numbers occur.

Every device communicates differently

One device outputs semicolons, the next outputs tab stops, and the third doesn't start the header until line 13. A separate Excel workaround is maintained for each format.

The assignment is done manually

Without a unique identifier, someone has to decide which sample a measurement belongs to. This works—until it doesn't, and it's noticed during an audit.

IT has to get involved with every connection

A new interface means a ticket, a service provider, and a wait. Until then, we'll continue using Excel for this intermediate step.

That is exactly what LDB LIMS’s import interface editor is for: a tool that allows you to integrate the many different devices, systems, and external laboratories—which vary widely in the frequency with which they provide data—ranging from an analyzer that reports hourly to a partner laboratory that sends a file three times a year.

Core Functions

Not just test results—almost everything that comes into the lab

The import interface editor is not limited to analytical instruments. Orders, master data, and documents are also imported through the same interface.

This data can be imported

  • Measured values
  • Samples
  • Orders
  • Customers and Contacts
  • Test Equipment
  • Materials and Batches
  • Attachments
  • Invoices and Quotes
  • Reports and Specifications
  • Formulas, Templates, Tasks

In addition, there are notes, time tracking, and files—a total of 19 data types that can be created or updated via import interfaces.

Supported Formats

CSVJSONXMLHL7LDTPDFPNGJPG

HL7 and LDT cover medical data exchange. Excel works as well—just save the file as a CSV in Excel. Everything else is added via a transformation code.

Transmission Paths

Which approach is right for your processes?

Four options, ranging from manual operation to fully automated. The decision depends on three factors: what your source system can do, what your IT infrastructure allows, and how often data is received.

Upload File

Manual

Upload the device's export file directly to the interface. The best way to get started—and for devices that are rarely in use.

IT effort: none
Good for: testing phase, small quantities

SFTP / FTP

Standard for Devices

The device saves the file on the server, and the lab database retrieves it automatically. Imported files are moved or deleted.

Retrieval: every 10 minutes
Quota: 10 files per cycle
Port: 22 (SFTP) · 21 (FTP)

HTTPS (REST-API)

Push and Pull

The external system actively sends the data via POST, or the lab database retrieves it. The fastest way—almost in real time.

Login: Basic Auth or certificate
Response: 200 successful · 400 rejected
Suitable for: ERP, portals, third-party labs

Email Receipt

Without an IT project

The interface accepts attachments via email. This is useful for partner labs that do not offer any other options.

IT effort: minimal
Good for: external labs, individual test results, temperature loggers

MCP

AI and Voice Assistants

ChatGPT, Claude, and other AI applications transmit data via open import interfaces—pathology findings are dictated rather than typed, sample collection is recorded via voice commands while on the go, and microbiology test results are documented hands-free.

Authorization: You determine which interfaces the AI is allowed to use
Rights: Linked to an LDB user
Audit Trail: Every change is marked with “MCP”
More about MCP

Custom Form

Capture Instead of Export

You can create your own HTML page or form—for on-site sampling, for submitters, or for a data collection process that doesn't yet exist. Once submitted, the entries go directly into the import interface.

Path: Form → HTTPS Submission
Good for: Samplers, submitters, special cases
Without a file: The entry is immediately saved as a record
Images and PDFs are assigned to the correct sample based on the file name, a QR code, or AI.
Audit Trail logs every import in full—including the raw data.
Import Mapping

Column by Field — Set it up once, and it runs on its own

During mapping, you specify which column in your file corresponds to which field in the lab database: Customer, Order, Sample, Sampling Location, Parameter, Measurement Value.

You upload a sample file, the lab database reads the attribute row, and you map the data. It doesn't matter whether the parameters are arranged side by side in columns or one below the other in rows—the mapping follows the structure of your file, not the other way around.

Screenshot: Import mapping with attribute mapping

Customize Label

Each attribute is given a descriptive name—a cryptic column heading becomes “Sampler.”

Set a default value

If a field in the import file is empty, the lab database automatically fills it with the stored default value.

Merge Fields

Multiple columns can be concatenated into a single field—first and last names become a single name.

Check Inputs

A template validates the data as soon as it is imported: If a value does not match the expected format, the import is aborted instead of creating invalid data.

What if the file isn't a CSV at all?

A transformation code then generates the required format before the mapping takes effect. Administrators in your lab can write and maintain this code themselves. Alternatively, we can handle this as a service at an hourly rate. In the import window, you will then see both side by side: the raw data and the converted version.

Transformation Code in the Manual
AI Extension

Convert unstructured data into structured data

Handwritten sampling logs, photographed measurement results, PDFs from clients or third-party laboratories: The AI extension converts unstructured documents into structured data, validates them, and passes them on to the import process.

Screenshot: AI-assisted import of a handwritten sampling log with validation

Two models instead of one

One model generates a result, and a second model independently verifies it. Both suggestions are displayed side by side, each with a confidence score—you can immediately see where they agree and where they don't.

Every value remains verifiable

For each field that is recognized, the lab database displays the location in the original document from which the value was derived. Any values that are ambiguous are flagged for manual review rather than being automatically accepted.

Access to Your Master Data

The AI is allowed to search the database for specific items—such as finding the testing equipment mentioned and correctly identifying it. It doesn't make recommendations; it simply looks things up.

For details on how the AI extension works, what data types it recognizes, and how to configure it for each interface, see the dedicated page.

View AI-powered imports
Import Mode

You decide what can be overwritten

An import can create, update, or deliberately leave everything unchanged. The mode is set for each interface—and then applies to every run.

The default is "update": The new value is always applied. In addition, there are modes that only fill in empty parameters, only update existing ones, create missing parameters without a value, or append measured values to an existing data series. If you do not want to overwrite any data during a repeat measurement, select an “ignore” mode—with or without a warning.

All Import Modes in the Manual
Screenshot: Selecting the import mode for importing measured values

Testing and Finding Bugs

Only in the test system

Each interface is first set up there and tested using anonymized sample files. It goes live only after successful validation.

Check Connection

One click tests SFTP or FTP: reachability, login credentials, and write permissions. The HTTPS interface returns a 200 or 400 response before any actual data is transmitted.

Test Special Cases as Well

Missing required fields, unknown IDs, special characters, empty values: Anything that goes wrong in everyday use belongs in the test—not in the first production run.

History and Repetition

Every import run is logged and can be repeated. If errors occur, you can see exactly where the problem lies—in the source file, the key, the mapping, or the transformation.

Devices & Systems

We have already integrated these devices and systems

And anything not listed here is added through mapping and transformation—the device's format simply determines how much preparatory work is required.

Analytical Instruments

  • Agilent
  • Bruker
  • Leco
  • Tecan
  • Zwick
  • GC/MS
  • HPLC · IC · MS
  • Waagen · Titratoren

Device Software

ERP & Inventory Management

  • SAP
  • SAP Business One
  • Microsoft Dynamics
  • Navision
  • ProAlpha
  • AS/400
  • Salesforce
  • DATEV

Practice & External Laboratories

  • Tomedo
  • M1
  • eTermin
  • Caritasnet
  • neolution
  • Novid20
  • HL7 · LDT
Can't find your device here? Send us a sample export file. In the interface workshop, we can usually determine within a single session whether mapping is sufficient or if a transformation code is needed.
Have the export file checked
Instructions

Import Interfaces in the Manual

Every step is documented—with screenshots, sample files, and videos. Freely accessible, even before you make a decision.

Basics

Creation & Mapping

Automation & Testing

AI & Special Cases

Practical Examples of Devices

Frequently Asked Questions About Import Interfaces and Device Connectivity

You can use import interfaces to connect analytical instruments, testing machines, temperature and data loggers, external laboratories, and ERP systems to the lab database. Possible data transfer methods include manual upload, SFTP, HTTPS/REST API, email reception, a custom form, or MCP for AI assistants. The appropriate method depends on the source system, your security requirements, and the desired level of automation.
Yes—and that surprises most people. Samples, orders, customers, and contacts; test equipment, materials, and batches; assets; invoices and quotes; reports; specifications; formulas; templates; tasks; notes; and time tracking are all managed through the same interface. There are 19 data types in total, and each one can be created or updated.
The most common formats are CSV, JSON, and XML. For medical data exchange, HL7 and LDT are also supported; in addition, PDF, PNG, and JPG files can be imported, such as chromatograms or signed reports. Excel also works—simply save the file as a CSV in Excel. When setting up the interface, specify the character set, date format, and decimal separator; incorrect settings are the most common cause of garbled umlauts or misinterpreted measurement values.
Yes—there’s no need for an intermediate step using Excel. The instrument output, parameters, sample ID, and data transfer path are configured once, and after that, the measured values are automatically assigned to the correct sample. This applies to instruments such as GC or PCR systems as well as to balances and titrators.
Yes, Zwick testing machines can be integrated via a compatible import interface. Depending on the configuration, we can transfer maximum values, average values, slopes, raw data, or entire curves—and you determine which of these measurement data points are actually transferred to the LIMS. Devices and test equipment are uniquely identified using fixed identifiers, serial numbers, or barcodes.
That’s where mobile data capture comes in: photos, data entry via iPad or iPhone, voice dictation, or traditional manual entry. Multi-page measurement reports and handwritten notes are scanned using OCR and AI and converted into structured measurement values. Sampling slips can also be scanned this way—which automatically generates orders and samples.
These include SAP, Navision, Microsoft Dynamics, and AS/400, among others. Transformation rules convert the ERP data into orders, samples, and assignments; the minimum data set usually consists of the customer number, item number, and batch number, with the unique assignment resulting from the combination of the batch and item numbers. Ideally, your ERP system remains the primary source for customer master data—the LIMS regularly synchronizes with it rather than creating a second source of truth. Invoice data is fed back via DATEV or a direct ERP interface.
Via the import interface with automatic parameter mapping or, for small quantities, by manual entry. The partner laboratory's original report can be attached to the sample so that the source of a value remains traceable during an audit.
Both transfer models are feasible. The frequency depends on the transfer method: scheduled retrievals, such as every ten minutes, or nearly instantaneous transfer via REST connections. Imports can also be triggered manually or based on a status value; status filters and unique import identifiers prevent duplicate imports and infinite loops.
Yes, via the Model Context Protocol (MCP). AI applications transfer data through only the import interfaces that you authorize for this purpose—they are not granted any further access. The login is linked to an LDB user; that user’s permissions apply, and every change is logged in the audit trail with the notation “MCP.” This allows a pathology report to be dictated rather than typed, or a sample collection to be recorded via voice while on the go.
Yes. You can create your own HTML page or form—for on-site sampling, for sample submitters, or for a data entry process that doesn’t yet exist in the LIMS. When submitted, the entries are sent directly to the import interface via HTTPS. No file is created that would need to be uploaded later.
Yes. In addition to individual uploads, there is a multi-upload option for entire folders—the standard method for importing legacy data. When retrieving data automatically via SFTP, the lab database fetches up to ten files per cycle, every ten minutes. Larger volumes are processed over multiple cycles or via an increased quota.
Yes. In addition to importing data into the lab database, data can be sent back to portals, ERP systems, or external labs—for results, status information, or IDs. The status that triggers the data transfer and the fields that are sent back are defined on a project-by-project basis.
During mapping, you map the fields in your import file to the fields in the LIMS: Customer, Order, Sample, Sampling Point, Parameter, Measurement Value. You can configure simple mappings directly in the interface. For complex file structures, non-standard labels, or special measurement value formats, a transformation code is required to convert the data before import. Administrators in your lab can write and maintain this code themselves; alternatively, we can provide this as a service at an hourly rate. Once created, the code should always be tested using representative test files.
Yes. AI recognizes columns, suggests mappings, and generates transformation code—allowing you to create an import mapping in just a few clicks instead of days. However, there are two rules to follow: anonymize sample files beforehand, and double-check every automatically generated mapping from both a business and technical perspective.
Both structures are supported. In a horizontal import, the parameters are arranged side by side in columns; in a vertical import, they are arranged one below the other in rows—device outputs are often vertical, while ERP exports are often horizontal. The only important thing is that the mapping matches the structure of the source file; you do not need to modify your existing exports to do this.
That depends on the import type. For measurement values, at least one valid sample, an assignable parameter, and the measurement value itself are required; for regulatory reports, IDs as well as operator or sampling site data are also required. Missing fields can be created in the system, and import templates can be configured differently for each location or test type.
Using identical item numbers and clean master data mapping. A test package can bundle multiple parameters that are collectively assigned to a single item number. Multiple batches from a single Excel file can also be automatically assigned and grouped together, provided they have unique characteristics.
Every data level requires a unique key: customer number, order number, sample number, parameter ID, patient ID, device ID, or barcode. If this key is missing, a wide range of tools are available to you—we use all available master data fields or combine multiple fields to create a matching key.
Yes. Barcodes identify samples, materials, batches, test equipment, and devices; materials and batches can be scanned directly. This requires that the code be stored consistently in both the source system and LDB LIMS and be included in the mapping. Alternatively, the serial number serves as a unique identifier for a device.
Through unique IDs, coordinated matching rules, and the appropriate import mode. The default “update” mode overwrites an existing value instead of creating a second record—if the same file is accidentally imported again, this prevents duplicate entries from being created. The definition of primary keys is particularly critical for customers, patients, locations, and samples. In addition, required-field and validation checks prevent incomplete or contradictory data from being processed in the first place.
Both can be configured on a field-by-field basis: You specify which fields an import is allowed to overwrite and which should remain unchanged. You should explicitly test empty values, existing content, and special business cases to ensure that nothing is overwritten unintentionally.
Yes. Each import run is assigned a unique ID, under which all associated data is stored and remains accessible at any time—including the raw data. This makes it possible to track when a particular file was imported, through which interface, which data records were generated from it, and who initiated the import. Each run can also be repeated.
Each interface is first set up in the test system and tested using representative sample files that are anonymized as much as possible. The testing covers not only standard cases but also, specifically, missing required fields, unknown IDs, special characters, empty values, and business-specific edge cases. Only after successful validation is the interface deployed to the production system—with separate login credentials and links for both environments.
First, check the source file, the keys used, the mapping, and the transformation logic; the error logs will show where the problem lies. The most helpful information for support is the anonymized import file, the affected interface, the sample or import ID, the specific error message, and a brief description of the result you expected. This allows us to pinpoint the cause much more quickly.
The creation and modification of import interfaces are restricted to administrative roles. Secure credentials are required for SFTP and API access, and separate test and production accounts are used; responsibilities among the laboratory, IT, and interface support teams are explicitly defined. Sensitive measurement and patient data are transmitted only via coordinated, secure channels; anonymization is mandatory for sample files and AI-supported mapping.
Yes. As part of the implementation, we use these exact import interfaces to migrate your legacy data—you don’t need a separate migration tool. This way, historical test results, customer and parameter master data, as well as large volumes of old calibration certificates—for example, from Lotus Notes—are transferred into the lab database and are then available for analysis as usual. Structured CSV files work best; the cleaner the structure of the legacy data, the less rework is required. The advantage: The interface you use for migration is the same one that continues to be used in day-to-day operations.
For a measurement data interface, the process works as follows:
  1. Open the device’s export file in Excel and save it as a CSV file with the UTF-8 character set.
  2. Check the attribute row and enter the parameter name from the export file in the " Import Key " field in the parameter master data.
  3. Under Samples → CSV Import Interfaces → New Interface, create the interface: select the “Measurement Data” type, set the character set to UTF-8, and check the “Sample Import” box.
  4. If the attribute row is missing or incomplete: Add column headers—they must be unique and may not contain spaces—and specify the row from which the import begins.
  5. Upload the file, perform the mapping, and assign the data.
  6. Navigate to the sample and upload the file there.
Afterward, test for errors and edge cases in the test system, go live with a small amount of data, and expand automation and monitoring. We’ll define required fields, status logic, device format, and matching rules together in the interface workshop.

Step-by-step guide with video: Import measurement data. Further reading: Import interface overview, create a CSV measurement data import interface, import results from measurement devices, transformation code for other data formats, automatic FTP/SFTP import, and AI enhancements for import interfaces.

Yes. Each import interface has its own HTTPS endpoint to which a third-party system sends its data via POST—authenticated using Basic Auth or a certificate. The endpoint returns a 200 status code if the data was accepted, and a 400 status code if it was received but could not be processed; each attempt appears in the interface overview. In the opposite direction, reports provide the data that a third-party system or an AI assistant is authorized to retrieve. The manual describes the structure and test calls.
Yes. Your administrators can create a new CSV interface themselves: upload a sample file, select the delimiter and character set, map the columns to the target fields, and save. Programming skills are only needed if a transformation code is required for an unusual format—and even that can be written and maintained by your administrators themselves.
Yes. You define the character set, date format, and decimal separator for each interface. A device that uses a period as the decimal separator or provides dates in the MM/DD/YYYY format will be read correctly as a result. Experience shows that incorrectly set values in this section are the most common cause of garbled umlauts and misinterpreted measurement values.
Yes, that’s actually the norm. Many labs start with the master data—customers, parameters, threshold values—and initially leave the results history in the legacy system. Any new data generated there during parallel operation can be imported later using the same import interface; a second migration run is not unusual. You decide how much historical data to transfer.
Yes, and the correction remains traceable. The audit trail records who changed which value and when; the originally imported value is not lost in the process. You use access rights to specify who is allowed to make corrections and who has read-only access.
This isn't folder monitoring in the traditional sense—SFTP is the way to do that. Your device or server stores the files in an SFTP directory, and the lab database automatically retrieves them every ten minutes—up to ten files per cycle—and then moves or deletes them. For your employees, the result is the same: simply saving the file is enough; no one has to upload anything.
Yes. The connection isn't just one-way: The lab database can send requests to the device or a third-party system via REST API—for example, it sends a sample number, and the device responds with the results for that specific sample. This is the cleanest approach if you don’t want the device to store any files or if you want to control when specific data is retrieved. You can use reports to specify which data the lab database should output.
Yes, both. We implement AnIML in collaboration with our partner Splashlake, and the lab database is SiLA-ready. These two standards are often requested in RFPs—but in practice, most integrations still rely on CSV, device-specific export files, and the REST API.

Onboarding: Import Interfaces

Automatically import measurements, samples and customer data from instruments, external labs and ERP systems

  1. Importing measurements from CSV files
  2. Access rights, import keys and import modes
  3. AI import mapping without programming skills
  4. Connecting GC and HL7 files via transformation code
  5. Automation via HTTPS and SFTP, integration with ERP systems
  6. Frequently Asked Questions on the TopicImport Interfaces and LIMS

LDB features

Laboratory Information Management System (LIMS)

QM modules

Simplify Documentation & Put ISO Requirements into Practice Every Day