Tagalog Phrases

Where this comes from

Every source, every licence, and who checked the Tagalog. This is a language site built by people who do not speak the language, so this page is the one that matters.

The frequency ordering is an estimate

This site's single substantive claim is that it teaches the phrases you are most likely to hear, in the order you are most likely to hear them. The honest version of that claim is: the ordering is an informed estimate, not a corpus count.

No published frequency count for conversational Tagalog exists. So every phrase carries source: estimate in the data, the site prints a rank and never a percentage, and if a real corpus count is ever built the order changes and the source field on every item changes with it.

Who has checked the Tagalog

Nothing is drilled until somebody has passed it. Here is exactly who has, and it is less than it should be:

Reviewed byDateScope
owner2026-08-2253 phrases
native speakerstill owedthe same 53 phrases, plus the 8 stress pairs

That was an owner's pass, not a native speaker's. The owner does not speak Tagalog. All 53 recordings were listened to by ear, and that caught a real defect: one came out as near-silence and now ships as a human recording instead. What a pass like that cannot catch is the failure the gate was actually built for. A phrase that is correct and that nobody says. A register a shade too formal for the situation it is taught in. A gloss that is accurate and is not what the phrase is for.

So reviewedBy travels with every phrase in the data. When a native speaker does go through them, what they changed and what had merely been assumed is answerable from the data rather than from memory.

Two things are deliberately not behind that gate: the 9-cell marker chart and the 21-cell pronoun grid. Those are settled facts of the language that appear in every reference grammar of it, they are not claims about what people say, and they are what makes the app teachable on day one. The 8 stress pairs are still shut, because the recordings that would make that drill possible do not exist yet.

The language data

The phrase list, the glosses and the cultural notes were written for this project. They are not extracted from a dictionary, a course or a copyrighted phrasebook. Where a fact about the language is not ours, which is to say the case-marker paradigm and the pronoun grid, it is a fact that appears in every reference grammar of Tagalog, and facts are not copyrightable. Schachter and Otanes, Tagalog Reference Grammar (1972), is the standard work and is what settles a dispute here. No text from it is reproduced.

The pronunciation transcriptions

The IPA on the 8 minimal pairs is checked against Wiktionary, whose text is CC BY-SA 3.0. A transcription of how a word is said is a fact rather than an expressive work, and none of its prose is reproduced. Which source each pair was checked against is recorded in the data, so a reviewer has something to check against.

The recordings

KindHow many
synthetic (fil-PH neural voice)153 files
human recording1 file

The synthetic recordings are Filipino neural voices, rendered with edge-tts. They are credited as synthetic in the app, because a learner is owed the knowledge that the thing they are copying is not a person. Human recordings replace them from the top of the frequency list down, which is also the order in which a synthetic voice does the most damage: the whole content of an interjection is its prosody.

Recordings that are not synthetic carry their own credit, rendered next to where the file plays rather than only here:

Type, code and artwork

The dictionary

The word lookup behind the phrase lookup is the Tagalog Wiktionary, 39,737 entries, extracted by kaikki.org (Tatu Ylonen) and served from this site's own database. Wiktionary content is CC BY-SA 4.0, which is why it is credited here and on every result it returns: the licence asks for attribution wherever the material is served, not merely wherever it was downloaded.

It is the only free source that marks Tagalog stress, which ordinary spelling omits and which changes what a word means. That is the reason it is worth carrying rather than a shorter word list.

It is a community dictionary, so it is broad rather than checked. Entries there have not been through the review this site's own phrases have, and the lookup keeps the two visually apart for that reason.

What this site measures

Two scripts, both served from this domain, both from Vercel, who host the site: Web Analytics counts page views and which link sent you, and Speed Insights times how fast the page painted. Neither sets a cookie, neither builds a profile, and neither can follow you to another website, so there is nothing here to consent to and no banner asking you to.

What it records is which of these pages people arrive on and which search sent them, so the ones nobody reads can be rewritten or dropped. It cannot see who you are.

Signing in is separate and optional. If you do, your progress is stored against an ID from Clerk -- crowns, XP, streak, and how often you got each cell right. There is no email address, no password and no page history in that table. Signed out, the app keeps everything in your own browser and sends nothing.

If something here is wrong

If you speak Tagalog and one of these phrases is a register too formal, or right in Manila and odd where you live, or simply not what anybody would say, that is the most valuable correction this project can receive. The review above stays open until a native speaker has been through all 53 of them.

Start with the free sectionThe six words that run the language — Tagalog puts a little word in front of every noun to show what it is doing. There are only six of them. No account, nothing to install.

More on this site