A free Japanese dictionary app feels simple until the verb in your sentence does not resemble the entry you expect. You read 食べました, 帰って or 話せなかった, but a conventional dictionary is organised around 食べる, 帰る and 話す. The form on the page carries tense, politeness, negation or another grammatical job; the headword removes those additions. If you search only the ending, guess the wrong word boundary or accept the first English definition, the lookup creates more confusion than it solves.
The practical answer is not to memorise every conjugation table before reading. It is to use a short recovery sequence: keep the original sentence visible, isolate the whole verb phrase, recognise the outermost ending, work back toward one or more plausible base forms, and test each candidate against the sentence. KanaWay's upcoming dictionary is designed to accept common conjugated forms and connect a useful result to saved review, but the reasoning below remains valuable with any reliable dictionary. It lets you judge the result instead of treating software as an answer machine.
A free Japanese dictionary app should reverse the form
Japanese dictionaries normally list verbs in the plain non-past form: 書く, to write; 食べる, to eat; する, to do; and 来る, くる, to come. This is commonly called the dictionary form because it is the form used for a dictionary headword. The Japan Foundation's teaching material makes that relationship explicit and shows how the three verb groups produce their dictionary forms. A lookup system therefore has to bridge two different objects: the surface form that performed a job in the sentence and the base entry that organises meanings and readings.
Treat that bridge as a hypothesis, not magic. 食べました strongly points to 食べる because the polite past sequence ました replaces the ending of an ichidan verb. Yet a short string may have more than one analysis, especially when kanji are absent or the sentence boundary is unclear. A good result shows the base form, reading, part of speech and detected inflection so you can inspect the route. Your job is to confirm that the candidate fits the surrounding particles, time reference and intended meaning before moving on.
Find the word boundary before changing the ending
The first mistake often happens before conjugation. Japanese normally has no spaces between words, and a verb can be followed by particles or helper expressions. In 本を読んでいます, the useful lookup is not んで or います alone. 読んで is the て-form of 読む, while います contributes the continuing-state construction. Copying the entire sentence may produce noise; copying too little removes the evidence that identifies the verb. Start by locating the final predicate and then expand left until you have the full lexical verb plus its grammatical tail.
Particles provide useful borders. In 昨日、駅で友だちを待ちました, the object marker を ends the noun phrase 友だちを and 待ちました forms the predicate. The adverb 昨日 supplies past-time context; 駅で marks the location. Isolating 待ちました gives the dictionary enough material to suggest 待つ without asking it to parse every noun. When you are unsure, keep two versions: the suspected verb alone and the complete sentence. Search the smaller unit, then use the larger one to verify meaning.
Peel off the outermost ending first
Long forms are easier when you reverse one layer at a time. Take 話せなかった, could not speak. The outer past-negative ending なかった points back to ない, giving 話せない. That form contains the potential ending せる attached to 話す. You do not need to jump from the full surface form to the answer in one guess. Each step asks a smaller question: what job does the final ending perform, and what form must have existed immediately before it? This also makes an incorrect analysis visible instead of hiding it inside a confident result.
Begin with high-frequency outer layers: ます and ました for polite non-past and past, ません and ませんでした for polite negatives, ない and なかった for plain negatives, and た or て families for completed or linking functions. Keep labels functional rather than translating each ending with one English word. The て-form, for example, supports requests, sequences, ongoing constructions and several other patterns. Identifying the shape narrows the lookup; the whole sentence tells you what that shape is doing here.
Reverse clear ichidan and irregular forms first
Group 2, often called ichidan, gives beginners the cleanest recovery pattern. Remove る from the dictionary form and add the required ending: 食べる becomes 食べます, 食べて, 食べない and 食べた. Working backward, 食べました loses ました and restores る; 見なかった loses なかった and restores る to become 見る. The method is powerful, but not every verb ending in る belongs to this group. 帰る, to return, and 走る, to run, are Group 1 verbs, so the dictionary candidate still needs verification.
Group 3 contains the two central irregular patterns. する produces します, して, しない and した; 来る, くる, produces 来ます, きます, 来て, きて, 来ない, こない, and 来た, きた. Compound verbs such as 勉強する follow する: 勉強して leads back to 勉強する. Learn these as compact families rather than exceptions scattered through a huge chart. They occur frequently, and resolving them quickly removes many false lookup paths before you tackle the more varied Group 1 endings.
Use the final sound to recover Group 1 verbs
Group 1, or godan, changes the final kana according to the form. In the て-family, 書く becomes 書いて, 泳ぐ becomes 泳いで, 話す becomes 話して, 待つ becomes 待って, and 読む becomes 読んで. The Japan Foundation's Irodori reference groups the changes usefully: う, つ and る move to って; ぶ, む and ぬ move to んで; く moves to いて; ぐ to いで; and す to して. 行く is the familiar exception, producing 行って rather than 行いて.
Backward lookup reverses those families, but sometimes produces several possible endings. A surface って may lead to a dictionary form ending in う, つ or る. 待って belongs to 待つ; 買って belongs to 買う; 帰って belongs to 帰る. Keep the kanji when available because it sharply reduces ambiguity. Without kanji, let the sentence select among readings and meanings. The aim is not to guess a kana ending in isolation; it is to find the base word that makes the complete sentence coherent.
Keep several candidates until the sentence chooses
Automatic deconjugation can return more than one technically possible base form. That is useful, not a failure. Consider かえって without kanji. Depending on the source, you may be looking at 帰って, the て-form of 帰る, or the adverb かえって, meaning instead or on the contrary. A dictionary can list both; only syntax and meaning can decide. If a candidate is a verb, check what particles precede it and what construction follows. If it is an adverb, check whether it modifies the wider statement rather than naming an action.
Do not choose by the first English gloss. Read the part of speech, kana reading and one concise example, then substitute the meaning into the original sentence. If the sentence still feels incoherent, return to the boundary or keep the next candidate alive. This is especially important with kana-only messages, subtitles and speech transcripts, where homophones share spelling. The fastest workflow is often ten seconds of structured comparison, not several minutes exploring every sense inside the first entry.
Use KanaWay as a lookup loop, not a reading detour
KanaWay's upcoming Japanese dictionary is designed around the form you actually encounter. Enter common conjugated verbs in kanji or kana and follow the detected form toward a likely base entry. Then compare its reading and common senses with the sentence you kept visible. The dictionary is intentionally compact rather than an exhaustive specialist reference, so it prioritises ordinary learner lookups and a quick return to the lesson, graded reading or listening task that created the question.
Save selectively. A word deserves review when it recurs, matters to your goals or blocks the current level—not simply because you looked it up once. Keep the base form and a trace of the original context, then let spaced review ask for recognition or recall later. Because dictionary search and saved words are designed to work offline, the same loop can continue on a commute. The app assists the transition from surface form to entry and from entry to review; your sentence remains the evidence that the result is right.
Return to the source and reread without the dictionary
A lookup is complete only after the original sentence works. Close or minimise the entry, reread the full line and state its main event in plain language. Then notice what the conjugation contributes: past time, negation, politeness, possibility, sequence or another relationship. For 昨日、駅で友だちを待ちました, knowing 待つ means to wait is only half the result. You should also see that 待ちました presents the waiting politely as completed in the past and that the other phrases establish when, where and for whom it happened.
If the sentence still fails, diagnose the remaining gap. Perhaps another word is unknown, the particle relationship is unclear or the dictionary sense was too broad. Make at most one additional lookup before rereading again. This limit protects the reading task from turning into open-ended reference browsing. For longer texts, mark nonessential unknown words and continue to the end of the paragraph before deciding whether they need attention. The dictionary should reduce interruption, not make every unfamiliar detail equally urgent.
Build a seven-day free Japanese verb lookup drill
For one week, collect one conjugated verb a day from material at your level. On days one and two, use polite forms such as 読みました and 食べません. On days three and four, use て-forms from short instructions or dialogues. On day five, choose a plain negative; on day six, choose a longer form such as 話せなかった. Before searching, circle the boundary, name the outer ending and write one likely base form. Then use a free Japanese dictionary app or trusted reference to check the path and sentence meaning.
On day seven, review the six original sentences with the answers hidden. Recover the headword, then explain what changed between it and the form in context. Count a success only when both the base verb and the grammatical job are clear. If one ending family repeatedly fails, make that the next week's narrow practice target. This small routine turns dictionary use into grammatical noticing without postponing real reading. More importantly, it trains the habit every future lookup needs: preserve context, analyse modestly, verify carefully and return to Japanese.
The best lookup is not the one with the most definitions. It is the one that explains the form, solves the sentence and lets you continue.
