Saral ↔ Unicode Gujarati Converter

This tool now handles both Gujarati conversions that used to sit on separate pages: Saral → Unicode and Unicode → Saral. By default the left box takes Saral ASCII text and the right box returns Unicode Gujarati. Going the other way needs no second page: the Swap button between the panes exchanges input and output, so you can paste Unicode Gujarati, swap it into position, and run the reverse pass from the same screen.

Both directions run on a single 111-entry mapping table compiled from the SaralSoft TTF cmap. The forward pass applies the table as-is; the reverse pass inverts it, longest match first, so a two-keystroke sequence like Aa (which renders as આ in Saral) resolves before its one-letter prefix A (અ). One table, two passes; there is no separately maintained reverse engine that can drift out of sync.

Nothing you paste leaves your device. The conversion is a few kilobytes of JavaScript running in your browser tab: no upload, no server call, no sign-up. A full 5,000-character passage converts in under 100 ms on a budget phone.

Saral is a legacy ASCII Gujarati font family from the pre-Unicode DTP era. It draws Gujarati letterforms on top of Latin character codes: with Saral applied, the keystroke k renders as ક and the four keystrokes wart render as ભારત. Underneath, the file stores plain English letters. Only the font makes them look Gujarati.

Unicode Gujarati is the modern standard. Every letter owns a dedicated code point in the U+0A80–U+0AFF block (ક is U+0A95), so the text displays identically on every phone, browser, and operating system with no special font installed.

Our mapping is ported from the open-source SaralSoft TTF cmap and covers Saral, LMG-Saral, LMG-Saral2, and OpenLMG-Saral. The four families share roughly 95% of their encoding, which is why one table serves all of them.

The forward conversion runs in two passes. Pass one is a regex that moves each pre-positioned short i-matra to after its consonant: in Saral you type િ before the letter it attaches to, while Unicode stores it after. Pass two substitutes the 111 mapping entries longest-first, parking every match in a private-use placeholder so already-converted output can never be re-matched by a later rule. That placeholder step matters for Saral in particular, because several Saral keys are ordinary digits and punctuation that would otherwise collide mid-conversion.

The workflow: paste Saral text on the left (it looks like scrambled English unless the font is installed; that is normal), let auto-convert fire or press Convert, then Copy output. The Unicode result drops straight into Word, Gmail, a CMS, or a government portal.

The reverse direction inverts the same table. Each Unicode Gujarati character or sequence maps back to the ASCII keystrokes Saral expects: ભારત becomes wart, આ becomes Aa, and the Gujarati numerals ૦ through ૯ become the accented capitals È through Ñ. Longest sequences match first, so conjunct forms resolve before the single letters inside them.

Expect the output box to look like gibberish. It should; those ASCII codes only turn back into Gujarati once you paste them into a document and set the font to Saral or one of its LMG siblings. Saral-era software understands nothing else, which is the whole reason this direction exists.

The workflow: place your Unicode Gujarati in the input box (press Swap first if it is sitting on the output side), then convert and copy the ASCII result. In your DTP file or document, paste it, select it, and apply the Saral font. The Gujarati reappears, glyph for glyph.

Saral → Unicode: old files, modern destinations

Gujarat's DTP shops, coaching centres, and government offices typed in Saral-family fonts for the better part of two decades. The result is an archive problem: PageMaker and CorelDRAW jobs, pre-2010 Word files, and coaching notes that are unreadable anywhere the font is missing. Converting to Unicode makes that material publishable on the web, searchable, and shareable over WhatsApp or email without attaching a font file.

For exam aspirants the direction is one-way traffic. GPSC and GSSSB typing assessments run on Unicode Gujarati interfaces, so Saral-era practice passages from an older institute need a forward pass before they are usable on a test screen. Our GPSC Gujarati typing test takes Unicode passages directly.

Unicode → Saral: feeding systems that never upgraded

Print and DTP workflows are the main customers here. A press running PageMaker 7 templates (letterheads, invitation cards, society notices) wants Saral ASCII it can drop into an existing layout, not Unicode it would have to retype. Reverse-converting your Unicode draft saves the operator the retyping, and saves you the proofreading of someone else's retyping.

The second audience is typing institutes that still teach on the Saral keyboard layout. Any Unicode article becomes drill material with one reverse pass. Offices holding decade-old Saral templates for certificates and registers use the same trick to keep those templates alive without re-setting them.

A font is a lookup table from character codes to drawings. Saral repurposes that table: the code for Latin k (U+006B) is drawn as ક, g as ગ, h as હ. Your file genuinely contains English characters; the font does all the Gujarati work at paint time.

The number row is letters too. Pressing 2 in Saral gives ધ, 3 gives ટ, 4 gives થ, 7 gives છ, 8 gives ષ, and 9 gives the conjunct ક્ષ. The Gujarati numerals ૦–૯ live elsewhere, on the accented Latin codes È through Ñ (U+00C8–U+00D1). Digit keys that type letters and letters that type digits: this one design decision causes most of the surprises listed in the troubleshooting section below.

Capital letters carry the half forms used in conjuncts. K types ક્, T types ત્, M types મ્, so the keystrokes Tya produce ત્યા. A handful of marks ride on punctuation: the full-stop key types the anusvar ં, the colon types the visarga ઃ, and the sentence-ending mark itself sits on the backslash.

The short i-matra follows the old typewriter convention: typed before its consonant, displayed before it. The keystrokes ipta render as પિતા. Unicode reverses the storage order (consonant first, matra after), which is exactly what the converter's regex pre-pass fixes for you.

All of this explains why Saral text shatters the moment it leaves a Saral-fonted environment. Paste it into a browser, open it on a phone, mail it to anyone without the font, and the disguise collapses into raw ASCII: ભારત turns back into wart. Unicode text cannot break this way, because the Gujarati lives in the characters themselves rather than in the font.

• Real numbers inside Saral text convert into letters. Saral treats 0–9 as letter keys (0 is એ, 1 is ઋ, 4 is થ, 9 is ક્ષ), so a phone number or a year embedded in Saral input comes out as Gujarati letters. Convert such lines separately, or restore the figures by hand afterwards. One quirk helps you spot trouble: 5 has no glyph in the standard Saral cmap and passes through untouched, so a lone 5 surrounded by Gujarati letters usually means the source mixed real numerals into Saral text.
• Gujarati numerals come out as accented Latin in reverse mode. ૦–૯ map to È É Ê Ë Ì Í Î Ï Ð Ñ. That is correct output, not corruption: apply the Saral font and the digits reappear.
• Full stops change glyphs. Saral draws its sentence-ending mark on the backslash key. A Unicode full stop becomes \ in reverse output, and a Saral \ becomes a full stop going forward. Both are by design.
• Short i-matra placement in reverse output. The reverse pass writes િ after its consonant (pita), whereas a hand-typed Saral file places it before (ipta). If your DTP software draws the matra on the wrong side of the letter, move the i one position left in the affected word.
• High-bit characters can be damaged before they reach the converter. ઑ, ૅ, ૉ, ઓ, ઔ, ઙ, and ઞ sit on Windows-1252 codes such as Š, š, †, ˆ, and ƒ. Copying through PDFs or old mail clients often substitutes or drops these bytes; if those letters vanish from your output, the input arrived damaged; re-copy from the original file rather than from the PDF.
• LMG variants differ on roughly 5% of glyphs. Saral, LMG-Saral, LMG-Saral2, and OpenLMG-Saral agree on the core mapping but diverge on a few rare conjuncts. Proofread anything high-stakes: a notice going to print, a legal document, an exam application.

• Convert in chunks. Up to 5,000 characters per pass works best. Long pastes from PDFs often carry hidden formatting that confuses any converter; break the text into paragraphs.
• Proofread conjuncts in both directions. Half forms ride on capital letters (K is ક્, T is ત્), and ક્ષ and જ્ઞ have dedicated keys. Skim the output for visibly broken stacks and repair them by hand.
• Diacritics convert reliably. Anusvar (ં), candrabindu (ઁ), and visarga (ઃ) all have fixed keys in the mapping. If your source uses a non-standard LMG variant, give these marks a second look.
• Check every figure. Digits are the most common casualty of any Saral conversion; see the troubleshooting list above before you trust a date, an amount, or a roll number.
• For exam answer scripts, skip the converter. Practise typing in the format your exam ships: Unicode Gujarati for GPSC and GSSSB (try our GPSC Gujarati test), Kruti Dev for SSC CHSL Hindi. A converter is a bridge for documents, not a substitute for finger memory.