About

Migrating contacts between systems (CRM, phone, email client) demands format fidelity. A malformed vCard file causes silent data loss: phone numbers assigned to wrong fields, truncated names, dropped email addresses. This converter parses CSV input compliant with RFC 4180 (handling quoted fields, embedded commas, escaped double-quotes) and produces vCard 3.0 output per RFC 6350. You control the column-to-field mapping explicitly. The tool validates EMAIL fields against RFC 5322 patterns and TEL fields for plausible digit counts. It does not guess. It does not auto-correct. You verify the preview table, then export.

Limitations: this tool processes structured CSV only. It will not parse fixed-width text or nested JSON. Binary photo embedding (Base64 PHOTO field) is not supported. vCard output uses UTF-8 encoding with line folding at 75 octets per spec. For files exceeding 10,000 rows, expect a brief processing delay. Pro tip: always check that your source CSV uses consistent quoting. A single unescaped quote character will shift every subsequent column.

Formulas

The converter maps each CSV row R_i to a vCard block. The mapping function M transforms column index j to vCard property P:

V_i = vCard(M(R_i)) = BEGIN:VCARD + n∑j=0 P_j:R_i,j + END:VCARD

Where V_i is the vCard output for row i, M is the user-defined column mapping, R_i,j is the cell value at row i column j, P_j is the vCard property assigned to column j, and n is the total number of mapped columns.

CSV field parsing follows RFC 4180. A field is quoted if it begins with ". Within a quoted field, a literal double-quote is escaped as "". The parser state machine uses three states: FIELD_START, QUOTED, and UNQUOTED. Transitions occur on encountering delimiter d (comma, semicolon, or tab), quote character ", or newline \n. Line folding in vCard output splits any property line exceeding 75 octets by inserting CRLF followed by a single space character.

Reference Data

vCard Field	Property Name	Example Value	Required	Common CSV Headers
Full Name	FN	Jane Doe	Yes	Name, Full Name, Display Name
Structured Name	N	Doe;Jane;;Dr.;	Yes	Last Name + First Name
Email	EMAIL	[email protected]	No	Email, E-mail, Email Address
Phone (Work)	TEL;TYPE=WORK	+1-555-0100	No	Work Phone, Business Phone
Phone (Cell)	TEL;TYPE=CELL	+1-555-0101	No	Mobile, Cell, Cell Phone
Phone (Home)	TEL;TYPE=HOME	+1-555-0102	No	Home Phone
Organization	ORG	Acme Corp	No	Company, Organization, Org
Title	TITLE	Engineer	No	Title, Job Title, Position
Address (Work)	ADR;TYPE=WORK	;;123 Main St;City;ST;12345;US	No	Address, Street, City, State, Zip
Address (Home)	ADR;TYPE=HOME	;;456 Oak Ave;Town;CA;90210;US	No	Home Address
Website	URL	https://example.com	No	Website, URL, Web
Birthday	BDAY	1990-05-15	No	Birthday, DOB, Birth Date
Note	NOTE	Met at conference 2024	No	Notes, Note, Comments
Nickname	NICKNAME	JD	No	Nickname, Alias
Role	ROLE	Project Manager	No	Role, Department Role
Categories	CATEGORIES	Client,VIP	No	Category, Group, Tags
Revision	REV	2024-01-15T10:30:00Z	Auto	Generated automatically
UID	UID	uuid-string	Auto	Generated automatically

Frequently Asked Questions

The parser implements RFC 4180 compliant quoted-field parsing. If a cell value is wrapped in double quotes, commas and newlines within are treated as literal characters, not delimiters. A double-quote inside a quoted field must be escaped as two consecutive double-quote characters (""). If your CSV file does not follow this convention, the parser will misalign columns. Check your export settings in the source application.

The output uses vCard version 3.0 (RFC 2426), which is the most widely supported version across devices and applications including iOS Contacts, Android, Outlook, and Thunderbird. Version 4.0 (RFC 6350) introduced minor property changes but has inconsistent client support. The converter adds VERSION:3.0 to every card and generates proper N (structured name) and FN (formatted name) properties, both mandatory in v3.0.

Yes. The converter provides separate mapping slots for Street, City, State/Region, Postal Code, and Country. These are assembled into the ADR property using the vCard structured value format: PO Box;Extended;Street;City;Region;Postal Code;Country. If any component column is unmapped, that position is left empty (double semicolon). Both WORK and HOME address types are supported independently.

Unmapped columns are silently ignored during export. They remain visible in the preview table but produce no output in the VCF file. This is intentional - CSV files often contain internal IDs, timestamps, or metadata irrelevant to contact data. No data is lost from the original file; only mapped fields are written.

Auto-detection performs case-insensitive fuzzy matching against a dictionary of 80+ common header variants. For example, headers like "First Name", "fname", "first_name", "Given Name" all resolve to the vCard N (given name) component. Email columns are validated against RFC 5322 simplified pattern. Phone columns are identified by headers plus digit-density analysis (>60% digit characters). The algorithm assigns confidence scores and only maps when confidence exceeds 70%. You can override every auto-detected mapping.

There is no hard-coded limit. The tool processes files in the browser using chunked parsing. Files under 5,000 contacts convert nearly instantly. Files between 5,000 and 50,000 contacts may take several seconds, with a progress indicator displayed. Beyond 50,000 rows, memory consumption depends on the browser. Chrome handles up to ~100,000 rows comfortably. For extremely large datasets, consider splitting the CSV file beforehand.

Yes. Email fields are checked against the pattern: local-part@domain with standard character restrictions. Phone fields are checked for a minimum of 7 digits after stripping formatting characters (+, -, spaces, parentheses). Invalid values are flagged with a warning icon in the preview table but are still included in the export - the tool warns but does not censor. You can review all warnings before downloading.