DIRECT SPEECH reproduces spoken words directly. English uses quotation marks for this, while Latin uses the defective verb "inquam".

(Pres. - inquam,inquis, inquit, inquimus, inquitis, inquiunt)
(Impf. - inquiebat)
(Perf. - inquii, inquisti)

Ea callida est, inquit.
(She is clever, he said.)

INDIRECT SPEECH reports spoken words indirectly, by means of an ACCUSATIVE AND INFINITIVE construction.

Dicit eam callidam esse.
(He says that she is clever.)

If the original discourse used a verb in the PAST tense, then a perfect infinitive is used.

Ea callida erat.
(She used to be clever.)

Dicit eam callidam fuisse.
(He says that she used to be clever.)

If the original verb was FUTURE tense, then a future infinitive is used.

Laeta erit.
(She will be glad.)

Dicit eam laetam futuram esse.
(He says that she will be glad.)

Dixit eam laetam futuram esse.
(He said that she was going to be glad.)