It usually suffices to call this template with its only required parameter, the class of the verb; for instance:

{{ta-conj-auto|3}}

This will create a Class 3 verb based on the current pagename. For example, if invoked on the தூங்கு (tūṅku, to sleep) page, it will produce

If a verb comprises of more than one word, only the final word will be conjugated.

|1=
The class of the verb. Required, and discussed below.
|2=
The verb, in the lemma (imperative) form. Normally, omit this parameter, and it will be derived from the page name, but it can be included (with the value of the verb’s infinitive) if it is not the same as the page name (e.g. in a test page in user space).
|imperat_sg=
The imperative-singular form, if it is not the same as the lemma
|imperat_pl=
The imperative-plural form, if it is irregular
|inf=
The infinitive form, if it is irregular. Notably, the irregular infinitives for வில் (vil), நில் (nil), கேள் (kēḷ), வா (), தா (), சா (), கல் (kal), கில் (kil), தோல் (tōl), and போ () are accounted for.
|adv_part=
The adverbial participle form when it is irregular, as in போய் (pōy) for போ ()
|aff_pres_base=
The (affective) present form, aka. the present adjectival participle for the verb if it is irregular. அழுகிற (aḻukiṟa) for அழு (aḻu), although this would be populated normally as அழு (aḻu) is a regular verb.
|aff_pres_avai_base=
Equivalent of the above, but in the third person plural neuter as the medial suffix often changes. அழுகின்ற (aḻukiṉṟa) for அழு (aḻu), although this would be populated normally as அழு (aḻu) is a regular verb.
|aff_past_base=
The (affective) past form, aka. the past adjectival participle for the verb if it is irregular. அழுத (aḻuta) for அழு (aḻu), although this would be populated normally as அழு (aḻu) is a regular verb.
|aff_futu_base=
The (affective) future form, aka. the base of the future forms (but not the future adjectival participle). அழுவ (aḻuva) for அழு (aḻu), although this would be populated normally as அழு (aḻu) is a regular verb.
|gerund_2=
The gerund (form II) when it is irregular, as in கேட்டல் (kēṭṭal) for கேள் (kēḷ)

An explanation of verb classes can be seen here or here. For reference, they're also described below.

Verbs that conjugate with (ta), கிற (kiṟa), and (va) in the past, present, and future, respectively. An example is அழு (aḻu, to cry)

Verbs that conjugate with ந்த (nta), கிற (kiṟa), and (va) in the past, present, and future, respectively. An example is உட்கார் (uṭkār, to sit)

Verbs that conjugate with இன் (iṉ), கிற (kiṟa), and (va) in the past, present, and future, respectively. An example is மூடு (mūṭu, to close)

Verbs that conjugate with final consonant doubling, கிற (kiṟa), and (va) in the past, present, and future, respectively. An example is சாப்பிடு (cāppiṭu, to eat)

Verbs that usually conjugate with (ṭa), கிற (kiṟa), and (pa) in the past, present, and future, respectively. All these verbs are irregular to some extent, and require the additional forms. An example is வில் (vil, to sell)

Verbs that conjugate with த்த (tta), க்கிற (kkiṟa), and ப்ப (ppa) in the past, present, and future, respectively. An example is படி (paṭi, to read, study)

Verbs that conjugate with ந்த (nta), க்கிற (kkiṟa), and ப்ப (ppa) in the past, present, and future, respectively. An example is நட (naṭa, to walk, happen)

Ordinary sandhi is taken into account, so excess forms for the present, past, etc should not need to be provided. For instance {{ta-conj-auto|1}} on the page செல் (cel, to go) produces

Notice how, by default, the future செல்லும் (cellum) includes the duplicated ல் (l) and the past tense base is சென்ற (ceṉṟa) due to natural sandhi between a lateral approximant and a dental or retroflex stop.