Convert youtube .vtt to .srt

vtt-to-srt.lua

local input_file = "tG1hCDXoE-I.en.vtt"  -- Change this to your VTT file path
local output_file = "output.srt"
-- Define input and output file paths.

-- Read the entire VTT file into an array of lines.
local lines = {}
for line in io.lines(input_file) do
    table.insert(lines, line)
end

-- Arrays for storing parsed values:
local timestamp_str_start = {}     -- Starting timestamp extracted from VTT (from line start to space after '>')
local timestamp_str_start_len = {} -- Length of the starting timestamp string (first element number value)
local timestamp_str_ending = {}    -- Ending timestamp snippet (the first non-blank line after three blank lines)
local top_subtitle = {}            -- Top subtitle (first line of caption)
local bottom_subtitle = {}         -- Bottom subtitle (second line of caption)
local timestamp = {}               -- Combined SRT-formatted timestamp (start --> end)

-- Initialize iterator for subtitle entries.
local i = 0
local line_index = 1
local first_timestamp_skipped = false -- Flag to track if the first timestamp has been skipped.

-- Loop through the VTT lines until the end of the file.
while line_index <= #lines do
    local line = lines[line_index]

    -- Detect a line with a timestamp (pattern: HH:MM:SS.mmm followed by whitespace)
    if string.match(line, "%d%d:%d%d:%d%d%.%d%d%d%s") then
        -- If the first timestamp has not been skipped yet, skip this one.
        if not first_timestamp_skipped then
            first_timestamp_skipped = true
        else
            -- Extract and store the starting timestamp.
            timestamp_str_start[i] = line:match("(%d%d:%d%d:%d%d%.%d%d%d)%s")
            -- Save the length of the starting timestamp as required.
            timestamp_str_start_len[i] = #timestamp_str_start[i]
            
            -- Read the following line as the top subtitle.
            line_index = line_index + 1
            top_subtitle[i] = lines[line_index] or ""
            
            -- Skip lines until three blank lines are encountered.
            local blank_count = 0
            while blank_count < 3 and line_index < #lines do
                line_index = line_index + 1
                if lines[line_index] == "" then
                    blank_count = blank_count + 1
                end
            end
            
            -- After three blank lines, read the next line for the ending timestamp snippet.
            line_index = line_index + 1
            if lines[line_index] then
                timestamp_str_ending[i] = lines[line_index]:match("(%d%d:%d%d:%d%d%.%d%d%d)")
            else
                timestamp_str_ending[i] = nil
            end

            -- Read the next line as the bottom subtitle.
            line_index = line_index + 1
            bottom_subtitle[i] = lines[line_index] or ""

            -- Combine the starting and ending timestamps into proper SRT format.
            -- Replace periods with commas.
            if timestamp_str_start[i] and timestamp_str_ending[i] then
                timestamp[i] = string.gsub(timestamp_str_start[i], "%.", ",") .. " --> " .. 
                               string.gsub(timestamp_str_ending[i], "%.", ",")
            end

            -- Increase the iterator to process the next subtitle block.
            i = i + 1
        end
    end

    -- Move to the next line in the file.
    line_index = line_index + 1
end

-- Write the parsed subtitles into the SRT output file.
local out = io.open(output_file, "w")
for j = 0, i - 1 do
    if timestamp[j] then
        -- Write the entry number.
        out:write(j + 1 .. "\n")
        -- Write the combined timestamp.
        out:write(timestamp[j] .. "\n")
        -- Write the top subtitle (first line of caption).
        out:write(top_subtitle[j] .. "\n")
        -- Write the bottom subtitle (second line) if it exists.
        if bottom_subtitle[j] ~= "" then
            out:write(bottom_subtitle[j] .. "\n")
        end
        -- Blank line separating entries.
        out:write("\n")
    end
end
out:close()

print("Conversion complete. Output saved as " .. output_file)