Devanagari Transliterator

How to Use the Devanagari Transliterator

1. 1. Choose a direction. Devanagari → Latin shows IAST and ISO 15919 side by side. Latin → Devanagari reads one system, which you pick, and writes the Devanagari.
2. 2. Paste your text. Hindi, Sanskrit or both. Conjuncts, nukta letters, anusvāra, candrabindu and visarga are read. Devanagari digits become 0–9, and the danda । becomes a full stop.
3. 3. Set the options. Choose how the anusvāra is written, whether to drop the final a of Hindi words, and (for ISO 15919) whether long e and o keep their macrons. Going back to Devanagari, choose what happens to a final consonant with no vowel.
4. 4. Copy the card you need. Each result says which system it is and who uses it. The letter table below the results lists every value.

IAST or ISO 15919?

Both systems give each Devanagari letter one Latin letter or digraph, so both can be reversed when the text is fully marked. They differ in a handful of places, and those are the ones you will notice in a reference list. IAST was fixed by Sanskrit scholars in 1894 and is what most Sanskrit editions and journals use. ISO 15919 (2001) was written to cover all the Indic scripts with one table, which is why it spells the vocalic r with a ring below and the long e and o with macrons. The Library of Congress Hindi table, which library catalogs follow, is a third, Hindi-specific variant.

The Rules the Tool Applies

• The inherent a. A consonant with no vowel sign is followed by a. A vowel sign replaces it, and a virama (्) removes it, so क्ष is kṣ and then takes its vowel. Both standards write the a everywhere, giving kamala, not the spoken kamal.
• Final a in Hindi. Hindi drops the final inherent a when speaking, so the option writes kamal and rām. It leaves single-syllable words such as न (na) and prefixes such as प्र (pra) alone, and it does not try to predict the silent a sounds in the middle of Hindi words.
• Anusvāra and candrabindu. The anusvāra is ṃ (IAST) or ṁ (ISO). Switch to the nasal option and it becomes ṅ, ñ, ṇ, n or m according to the next consonant, the rule in the Library of Congress Hindi table. A candrabindu is m̐ in IAST; ISO 15919 puts a tilde over the vowel (hãsī) and falls back to m̐ when there is no vowel to carry it.
• Conjuncts. ज्ञ is jñ and क्ष is kṣ, the Sanskrit readings. Hindi speakers say gy and chh, but transliteration records the letters, not the sound.
• Nukta letters. क़ ख़ ग़ ज़ झ़ फ़ य़ and the flaps ड़ ढ़ are written q, k͟h, ġ, z, ž, f, ẏ, ṛ and ṛh. IAST does not define them, so both columns use the ISO values.
• Separators. If a stop is followed by ह across a virama (तद्हित), the tool writes a hyphen (tad-hita) so dh is not mistaken for ध. A hyphen typed in Latin text keeps two letters apart on the way back.

Going Back to Devanagari

The reverse direction reads a Latin letter or digraph at a time, longest first, and builds the syllables: a consonant plus a vowel becomes a consonant with a vowel sign, a consonant plus another consonant gets a virama, and a adds nothing. That is exact for fully marked text such as the converter's own output. It has three limits.

• A final consonant is ambiguous. kamal might be कमल (Hindi) or कमल् (Sanskrit, with a virama). The option decides; the default is the Hindi form.
• Digraphs always win. kh is ख, never क followed by ह. Use a hyphen (k-h) when you mean two letters.
• ai and au are always diphthongs. The sequence अइ would be written ai and read back as ऐ. The rare cases are not marked in either standard.

Hindi and Sanskrit Names in Citations

A reference list gives Devanagari titles and author names in Latin letters so that readers who do not read the script can find the work. Choose one system and keep it for the whole list; Sanskrit and Indology journals expect IAST, and library records follow the Library of Congress table. How to Cite a Source in Another Language shows where the romanized title and its English translation go for APA, MLA and Chicago, and the Book Citation Generator formats the finished entry.

Frequently Asked Questions

What is the difference between IAST and ISO 15919?

They share most letters. IAST writes the vocalic r as ṛ, the long e and o as plain e and o, and the anusvāra as ṃ. ISO 15919 writes r̥, ē and ō (with plain e and o allowed for languages that have no short ones), ṁ for the anusvāra, and a tilde for the candrabindu. ISO 15919 also assigns ṛ to the Hindi flap ड़, which IAST leaves undefined.

How do I transliterate Hindi to English letters?

Paste the Hindi into the converter and pick the system your instructor or publisher names. For a name in running text, turn on Drop it — Hindi pronunciation to get Rām instead of Rāma, or use the spelling the person or place already has; established forms like Delhi or Krishna are not romanizations of any standard.

Why does the output show a at the end of every word?

Because the standards transliterate the written letters, and in Devanagari every consonant carries an inherent a unless a sign cancels it. Hindi speakers drop the last one, which is why कमल is spoken kamal. Turn on Drop it — Hindi pronunciation if you want the spoken form.

How do I type IAST letters like ā, ṣ and ṃ?

Use a keyboard layout with dead keys or your system's character map: ā ī ū ṛ ṝ ḷ ṭ ḍ ṇ ṅ ñ ś ṣ ṃ ḥ. You can also copy them from the converter's output or its letter table.

Is the ISO 15919 anusvāra ṁ or ṃ?

In the simplified form of the standard it is always ṁ (m with a dot above). The strict form writes the nasal of the next consonant, which the nasal option gives you. ṃ, with the dot below, is the IAST letter.

Why is ज्ञ written jñ and not gy?

Transliteration records the letters: ज, a virama and ञ. In Sanskrit the cluster is jñ; in Hindi it is pronounced gy (as in ज्ञान, gyān). A transcription of Hindi pronunciation would write gy; this tool, like both standards, writes jñ.

Does it work for Marathi, Nepali or other languages in Devanagari?

It will convert the letters, but those languages have their own romanization tables with different values for some letters, such as ळ. For a catalog record, use the table for that language.