Command-Line X12 EDI File Converter
Overview
Our command-line EDI conversion tool converts X12 EDI files to JSON or CSV/Excel. Conversion to CSV/Excel is supported for 837 (healthcare claim), 835 (ERA) and 834 EDI files.
The output is easy to understand and requires no EDI knowledge.
To familiarize yourself with the output formats, review our intro to the JSON format and the intro to the CSV format.
Output examples:
- 835: CSV, Excel, JSON
- 837P: CSV, Excel, JSON
- 837I: CSV, Excel, JSON
- 837D: JSON
- 277CA: JSON
- 834: CSV, Excel, JSON
For more examples, go to our list of sample EDI files, select any sample, and click export.
You can also use our free EDI Viewer to convert small files (up to three claims). Upload your file and export your data as CSV or JSON from the UI.
The CLI tool produces the exact same output as our API server, except for code descriptions.
See our JSON schema and CSV data dictionary documentation for the details about CSV columns and JSON fields.
The converter can efficiently handle large X12 EDI files with special support for 837/835 files. See this section for more details.
The conversion is performed locally on your machine; your data never leaves your network.
The converter’s only dependency is Java. Windows, Linux, and macOS are all supported.
Current release: 2.14.4. What’s new: Support for PLB (provider-level adjustments) in 835, MIA/MOA segments, 834 JSON and CSV, 277CA JSON.
Installation Steps
- Download and install Java from the Oracle website. You can also use OpenJDK. Java 17 or higher is required. We recommend installing the latest long-term support (LTS) release of Java. You can also install Java using sdkman or a package manager on Linux or macOS. Once you install Java verify the Java’s version by running
java -versioncommand from the command line window/terminal. - If Java is already installed on your system, verify that you have the correct version by running
java -versioncommand; it must be version 17 or higher. - If you have
JAVA_HOMEenvironment variable defined, make sure that it points to Java 17 or higher. Alternatively, you can delete theJAVA_HOMEvariable. To view environment variables on Windows, go to “Settings/Advanced system settings”, “Advanced” tab, click “Environment Variables…”. - Download the zip file with the converter from this link.
- Unzip the converter’s zip file; the root folder/directory of the installation will be
ediconvert.
Many Windows systems have Java 8 pre-installed. The converter requires Java 17 or later; it will fail when Java 8 is used. Follow the steps above to ensure that you have the correct version of Java.
You need a license key to run the converter. Request your trial license key here.
Once you receive your edi-license.txt license file, save it into the installation folder, i.e., into ediconvert. Alternatively, you can save the license file in a different location; then, you need to create the EDI_LICENSE_FILE environment variable that points to it. You can also copy the license key to the EDI_LICENSE_KEY environment variable.
You can run the converter by providing the full path, e.g., ediconvert\bin\ediconvert on Windows or ediconvert/bin/ediconvert on Linux/macOS.
We recommend adding the bin directory of the converter to the “Path” so you can run it from anywhere. On Windows, go to “Settings/Advanced system settings”, “Advanced” tab, click “Environment Variables…”, click a variable called “Path” and click “Edit…”. Create a new entry with the absolute path to the bin folder. On Linux/macOS, append it to the PATH variable in your shell configuration file, e.g., for .bashrc: export PATH=$PATH:ediconvert/bin.
All examples in this document assume that you’ve added the converter to the path, so we omit the full path to bin/ediconvert.
Run ediconvert -V to view the license and version information.
Running the Converter
To run the converter, provide a path to an EDI file and an output directory (folder) or a file name:
mkdir json
ediconvert edi/837/my_837_file.edi -o converted-files -m json
The -m (--mode) option defines the output format: JSON, line-delimited JSON/NDJSON (jsonl) CSV, or Excel. JSON is the default.
You can use wildcards to convert multiple files in one go (make sure to enclose the first parameter into quotes):
ediconvert "edi/837/*.edi" -o ./converted-files
Due to automatic wildcard expansion by command-line shells and the Java platform, always enclose a parameter with wildcards into double quotes on Linux/macOS or single quotes on Windows.
If you provide an existing directory/folder as the output, the converter will convert each file from the input into this directory. The converter will create output file names based on the input file names with the appropriate file extension depending on the output format.
If the output is not an existing directory, the converter will assume the file as the output, and it will convert all input files into one file:
# Convert all files from 835 folder into converted-files/combined-835.csv file
ediconvert "edi/835/*" -o converted-files/combined-835 -m csv
In other words, if you want to convert each input file into a separate output file, provide the location of an existing folder.
If you want to convert all input files into a single file, provider the name of the output file.
Convertion to CSV
When converting 835/837 files, the converter creates two output files in CSV format, one for claim-level data and one for line-level data. The line-level file has the “-Lines” suffix.
For example, ediconvert edi/837/my_837_file.edi -o converted-files -m csv creates my_837_file.csv and my_837_file-Lines.csv in the converted-files folder.
The internally generated ID, the patient control number, and the payer control number (for 835) fields are repeated for each line in the line-level file:
You can also ask the converter to produce a single file using the --single-csv option. In this case, the output file will contain both claim and line-level columns.
The claim-level data will be listed once, and ID and patient control number fields will be repeated for each line of the same claim:
All CSV columns and their mappings to X12 EDI are documented in our data dictionary.
The CSV output format is governed by a configuration file containing several pre-defined configurations (conversion schemas).
The conversion schemas define what columns to include in the CSV file and how many times to repeat columns for repeating groups, such as adjustments or diagnosis codes.
You can create your custom conversion schemas or select one of the built-in schemas by providing --csv-schema-name option.
For example, if you want to repeat all claim-level fields for all lines (as opposed to just the first one), you can use the schema lines-with-header-repeat-each-row:
ediconvert edi/837/my_837_file.edi -o converted-files -m csv --csv-schema-name lines-with-header-repeat-each-row
Please follow this document for more details.
Convertion to Excel
The converter can generate Excel instead of CSV. The CSV-related command-line options, such as --csv-schema-name and --single-csv apply to Excel:
ediconvert edi/837/my_837_file.edi -o converted-files -m excel --single-csv
However, instead of generating separate files for claim and line level data, the converter creates a single Excel file with different sheets.
The converter sets column types based on the type of the EDI field. For example, the “ChargeAmount” column is defined as numeric in the generated Excel file. This saves you time—instead of manually importing CSV and setting column types, you can generate the Excel file that can be immediately used for analysis.
Note that Excel files have limitations in terms of the number of rows, so converting to Excel is suitable only for relatively small datasets.
On macOS, you may see the following messages when converting to Excel; they can be ignored:
2025-06-19 13:01:19.776 java[24509:126725106] +[IMKClient subclass]: chose IMKClient_Modern
2025-06-19 13:01:19.776 java[24509:126725106] +[IMKInputSession subclass]: chose IMKInputSession_Modern
Converting Large Files and Large EDI Transactions
For JSON conversion, the converter supports NDJSON (a.k.a. “JSON Lines”) format with line-separated JSON objects. This provides an alternative to well-formed JSON with enclosing arrays, which are inefficient with large files.
Use the -m jsonl option to produce NDJSON files instead of well-formed JSON.
The converter automatically splits large 837/835/834 transactions. It does this by reading a batch of claims (CLM segment for 837), payments (CLP segment for 835), or members (INS segment for 834) instead of a batch of transactions.
The size of the batch is controlled by the --chunk-size option. The default value is 50 claims/payments/members for 837/835/834 transactions or 50 transactions for all other transaction types.
Increasing the chunk size can improve performance, but it can also increase memory usage.
Example:
ediconvert edi/837/edi_file_large_transaction.edi -m jsonl --chunk-size 100
Error Handling
If the converter cannot parse an EDI segment, it will produce a parsing warning. Warnings are logged and can optionally be added to the converter’s output (JSON or CSV) as WARNING objects. This behavior is controlled by --out-parsing-warnings command-line option.
For more details on the format of warning messages, see this page.
Command-line Options
Usage: ediconvert [-hqrV] [--format-json] [--out-parsing-warnings]
[--out-segments] [--single-csv] [--[no-]split-tran]
[--chunk-size=<chunkSize>]
[--csv-config=<csvConversionSchemaFileName>]
[--csv-schema-name=<csvConversionSchemaName>]
[-m=<outputFormat>] [--max-warnings=<maxWarnings>]
[-o=<outFile>] [-p=<filePatterns>]
[--progress-chunk-size=<loggingChunkSize>]
[--tran-type=<transactionTypeFilter>] EDI File or Directory
Converts X12 EDI files to JSON or CSV
EDI File or Directory
X12 EDI file or a directory (folder) containing X12
EDI files. It can contain '*' and '?' wildcards to
match multiple files. Enclose in single quotes on
Windows or '"' on Linux if using wildcards.
-p, --patterns=<filePatterns>
File name patterns to match in the directory, e.g.,
'*.edi'. You can specify multiple comma-delimited
patterns, e.g., '*837*.edi, *837*.txt'. Defaults to
'*' (all files) if 'EDI File or Directory' is a
directory. Enclose in single quotes on Windows or
'"' on Linux.
-r, --recurse Recursively search for matching files in all
subdirectories
-o, --out=<outFile> Output file or existing directory where to save
output files. The file will be created if it
doesn't exist. For a directory, the output file
name will be the EDI file name plus the extension,
depending on the output format. If omitted, the
converter will output to the console (standard
out), which can be piped or streamed.
-m, --mode=<outputFormat>
Output format. Acceptable values: CSV, EXCEL, JSON,
JSONL. Defaults to 'json'.
--format-json Format JSON with indentation and new lines. Ignored
for CSV and JSONL.
--out-parsing-warnings
Write EDI parsing warnings to the output file as
'warning' objects. If not set, parsing warnings are
written to the log and standard output. See https:
//datainsight.
health/docs/ediconvert-api/user-guide/#error-handlin
g for more details.
--max-warnings=<maxWarnings>
The maximum number of parsing warnings per file
before stopping and raising an error. Defaults to
100. Set to -1 to suppress raising the 'too many
warnings' error; the converter will process the
entire file.
--chunk-size=<chunkSize>
The number of EDI transactions or
claims/payments/members for 837/835/834 that will
be parsed and converted at once. This option
determines the memory footprint and can be used for
fine-tuning. Defaults to 50. Higher value improves
performance but consumes more memory.
--[no-]split-tran Do not parse the entire transaction. Instead, parse
by claims (CLM or CLP segments, 837/835) or members
(INS segment, 834). This is the default for 835,
837, and 834 transactions, which can be large.
Ignored for all other transaction types. Defaults
to 'true'
--progress-chunk-size=<loggingChunkSize>
How often to log progress. By default, the conversion
progress is logged every 1000 items (claims,
members, transactions)
-q, --quiet, --suppress-progress
Suppress logging file names and conversion progress
to the console; use this option when converting to
standard out
--out-segments Convert to the X12 EDI loop and segment tree instead
of the list of claims, payments, and other objects.
JSON or JSONL only. It can be slow for large
transactions.
-h, --help Show this help message and exit.
-V, --version Print version information and exit.
Options for CSV conversion:
--csv-schema-name=<csvConversionSchemaName>
Name of the 'schema' (configuration) in the CSV
config file, such as 'key-fields'. The schema
determines the layout of the CSV file. See https:
//datainsight.health/docs/csv/schemas for the list
of schemas.
--csv-config, --csv-schema-file=<csvConversionSchemaFileName>
Name of the file containing CSV conversion schemas in
the 'conf' directory. Defaults to
'conf/csv_conversion.yaml'.
--tran-type=<transactionTypeFilter>
Convert only the files that contain transactions of
this type. All other files will be ignored.
Acceptable values: 835, 837P, 837I, 837D, 834.
--single-csv Convert all EDI files into a single CSV file
containing the claim-level and the service
line-level fields (837/835). If not provided, the
converter will convert claim-level and line-level
data into separate files. Same as providing
'--csv-schema-name=single-file'. A single file is
the default for 834.
Converting to the Console (Standard Output Stream)
You can run the converter in the JSON Lines or in the CSV mode and pipe its output to another process, e.g.:
ediconvert edi/837/my_edi_file.edi -m jsonl -q | grep '"zipCode":"37232"'
Make sure to use -q option when converting to the console.
Healthcare and X12 EDI Codes
The CLI tool does not provide descriptions for healthcare codes, such as ICD-10 and CPT. Use our API server product if you’re interested in getting descriptions along with the codes.
You can also use our code lookup API to programmatically obtain descriptions for healthcare codes. Request your API key here.
The CLI converter converts some of the EDI codes, such as an “Entity Identifier code” to mnemonic constants, e.g., the code 77 is translated to SERVICE_FACILIY.
You can always find the code for the constant using the code lookup tool. Enter a name of any constant to see its description and the EDI code it corresponds to, e.g., SERVICE_FACILITY.
License File and License Key Configuration
The default location of the license file is the converter’s installation folder, e.g., ediconvert.
Alternatively, you can save the license file anywhere and define the environment variable EDI_LICENSE_FILE with the path to the license file:
For Linux/macOS, add the variable to your profile:
export EDI_LICENSE_FILE=<absolute path to the file>
For Windows, create a new environment variable EDI_LICENSE_FILE from “Settings/Advanced system settings.”
You can also set your license key using the environment variable EDI_LICENSE_KEY. Copy the entire content of edi-license.txt into that variable. In this case, you don’t need to save the license key file anywhere.
Run ediconvert -V to view the license information and the location of the license file.
Logging
The converter creates logs with information and diagnostic messages in the “logs” directory. These logs are useful when routing the output to standard out.
The converter always logs to the file, even when the --quiet option was set.
The default location of the logs folder is the directory where the converter is executed. It can be changed by defining the ediconvert_logs_dir environment variable, e.g.,
export ediconvert_logs_dir=ediconverter_logs